Abdominal girth and vertebral column length can adjust spinal anesthesia for lower limb surgery, a prospective, observational study

Background Studies have shown that abdominal girth and vertebral column length have high predictive value for spinal spread after administering a dose of plain bupivacaine. we designed a study to identify the specific correlations between abdominal girth, vertebral column length and a 0.5 % dosage of plain bupivacaine, which should provide a minimum upper block level (T12) and a suitable upper block level (T10) for lower limb surgeries. Methods A suitable dose of 0.5 % plain bupivacaine was administered intrathecally between the L3 and L4 vertebrae for lower limb surgeries. If the upper cephalad spread of the patient by loss of pinprick discrimination was T12 or T10, the patient was enrolled in this study. Five patient variables and intrathecal plain bupivacaine dose were recorded. Linear regression and multiple regression analyses were performed. Results Totals of 111 patients and 121 patients who lost pinprick discrimination at T12 and T10, respectively, were analyzed in this study. Linear regression analysis showed that only abdominal girth and plain bupivacaine dose were strongly correlated (r =−0.827 for T12, r = −0.806 for T10; both p < 0.0001). Multiple linear regression analysis showed that both abdominal girth and vertebral column length were the key determinants of plain bupivacaine dose (both p < 0.0001). R2 was 0.874 and 0.860 for the loss of pinprick discrimination at T12 and T10, respectively. Conclusions Our data indicated that vertebral column length and abdominal girth were strongly correlated with the dosage of intrathecal plain bupivacaine for the loss of pinprick discrimination at T12 and T10. The two regression equations were YT12 = 3.547 + 0.045X1-0.044X2 and YT10 = 3.848 + 0.047X1- 0.046X2 (Y, 0.5 % plain bupivacaine volume; X1, vertebral column length;and X 2, abdominal girth), which can accurately predict the minimum and suitable intrathecal bupivacaine dose for lower limb surgery to a great extent, separately.


Background
Plain bupivacaine spinal anesthesia is a conventional technique [1][2][3][4][5]. However, the spread of intrathecal plain bupivacaine is highly unpredictable [6]. Previous studies have proven that at least 25 factors could affect the spread of spinal anesthesia [7][8][9]. Lumbosacral cerebrospinal fluid volume was found to be the primary determinant that affect spinal cephalad spread [10], lumbosacral cerebrospinal fluid pressure was also an important factor that affects spinal cephalad spread [11], but those informations had little practical value because they were inconvenient to obtain. Clinically, patient's age, height, weight, and body mass index (BMI), have commonly been considered when determining the dose of bupivacaine. However, previous studies have reported that these variables were only weakly to moderately correlated with and had little predictive value for the spread of spinal anesthesia [1,3,[12][13][14].
Our previous study showed that abdominal girth and vertebral column length were highly predictive of the spread of spinal anesthesia after a given dose of plain bupivacaine [15]. However, the results did not suggest a specific dose of bupivacaine for an individual. T 12 spinal spread level was nearly the minimum block level for eliminating tourniquet pain and T 10 spinal spread levels could provide satisfactory analgesia with almost no hemodynamic effects for lower limb surgery. Plain 0.5 % bupivacaine has frequently been used for spinal anesthesia [1,3,13,16]. Therefore, we designed a prospective, observational study to identify the specific correlations between abdominal girth, vertebral column length and a 0.5 % dosage of plain bupivacaine, which should provide a minimum upper block level (T 12 ) and a suitable upper block level (T 10 ) for lower limb surgeries.

Ethics
This study was approved by the Ethical Committee of the Second Affiliated Hospital of Jiaxing University (Ethical Committee number CZJE 201311; Chairperson Professor Li-qin Jiang) on 11 February 2013. All patients provided prior written informed consent.

Patients
A total of 692 patients were enrolled in this study from February 2013 to March 2014, and the following inclusion criteria were applied: ASA physical status I and II, 19-60 years old, and were scheduled for lower limb fractures surgery, hip arthroplasty, knee arthroplasty, knee arthroscopy, great saphenous vein surgery and so on. Patients with contraindications for spinal anesthesia, diabetes, rheumatoid arthritis, spinal canal stenosis, a history of spinal anesthesia or spinal surgery, pregnant patients and patients with language disorders were excluded.

Intervention
All patients fasted for 8-10 h before surgery. After the patients entered the operating room, intravenous access was established, and Ringer's lactate 10 ml/kg was preloaded before anesthesia. Noninvasive arterial pressure was recorded every 5 min; electrocardiography and pulse oximetry were continuously monitored for all patients. The patients were placed in supine position, and at the level of the umbilicus, the abdominal girth was measured at the end of expiration. Before spinal anesthesia was administered, the patient was placed in the right lateral decubitus position, with spinal column flexion. During spinal anesthesia, the L3-4 interspace was confirmed by ultrasound imaging, a 25-gauge Quincke spinal needle was inserted, and a midline approach was adopted, with the orifice pointing toward the cephalad. We set T 10 spinal block level as the target level, and a dose of 0.5 % plain bupivacaine that the anesthesiologist deemed suitable was injected intrathecally with room temperature at a speed of approximately 2 ml in 10 s when free flow of the cerebrospinal fluid was obtained. After the procedure, the patient was placed supine position immediately. The spinal anesthesia spread was assessed using an 18-gauge sharp needle every 10 min in both midclavicular lines till 60 min after the intrathecal injection. At each level, three pin touches were employed for the loss of pinprick discrimination. If the ultimate spinal anesthesia level of loss of pinprick discrimination in both midclavicular lines was T 12 or T 10 in 60 min after the intrathecal injection, the patient was picked out to analyze in our study, and the vertebral column length was measured from the C7 vertebra to the sacral hiatus, with the patient placed supine on the operation table. The sacral hiatus and C7 vertebra were both confirmed by radiographic imaging.
Two anesthetists participated in the study. All anesthesia procedures and patient management were performed by the same anesthetist, and assessment of the spinal anesthesia spreads were performed by another anesthetist, who was unaware of the bupivacaine dose used in each patient. Patients , age, height, weight, vertebral column length abdominal girth and 0.5 % bupivacaine dose were recorded.

Statistical analysis
The sample size was determined as described in reference 15. If the anticipated effect size was 0.15, the desired statistical power level was 0.8, the predictors were 5, and the minimum required sample size was 91, with a probability level of 0.05. SPSS software (SPSS Inc., Chicago, IL, USA), version 19.0, was used to analyze the data. Linear regression analysis was performed to determine the correlations between the spinal anesthesia -induced loss of pinprick discrimination at T 12 or T 10 and patients , age, weight, height, abdominal girth, and vertebral column length. Multiple regression analysis was performed to test the combined linear contributions of patients , age, height, weight, vertebral column length and abdominal girth to the bupivacaine dose for T 10 or T 12 loss of pinprick discrimination. During multiple regression analysis, stepwise selection was performed, and the predictors were removed from the analysis if they were not significantly correlated with the dependent variable. R 2 was the coefficient of determination. A p value < 0.05 was considered to be statistically significant.

Results
Of 692 potential participants, 45 failed to meet the inclusion criteria,391 patients had spread levels other than T 10 or T 12 (the maximum block level was T 4 , the minimal block level was L 2 ), and 24 were excluded for other reasons. The remaining 232 patients (111 patients with loss of pinprick discrimination at T 12 , 121 patients with loss of pinprick discrimination at T 10 ) were analysed in this study (Fig. 1). There were no significant differences in the variables between those patients with loss of pinprick discrimination at T 12 and at T 10 (p > 0.05) ( Table 1).
Linear regression analysis showed that there were significant univariate correlations between height, weight, abdominal girth, and vertebral column length and the 0.5 % plain bupivacaine volume for loss of pinprick discrimination at T 12 or T 10 (all p < 0.0324). However, only abdominal girth and 0.5 % plain bupivacaine volume were strongly correlated (r = −0.827 for T 12, r = −0.806 for T 10; p < 0.000) ( Table 2; Figs. 2 and 3).
Multiple linear regression analysis showed that abdominal girth and vertebral column length were the key determinants of 0.5 % plain bupivacaine volume (both p < 0.0001), while age, weight, and height could be omitted without changing the results (all p > 0.164, all 95 % confidence intervals < 0.267) ( Table 3). R 2 was 0.874 for the loss of pinprick discrimination at T 12 and 0.860 for the loss of pinprick discrimination at T 10 .

Discussion
The most important finding of the present study was that vertebral column length and abdominal girth were strongly correlated with the dosage of intrathecal plain bupivacaine for the loss of pinprick discrimination at T 12 and T 10 . This finding indicated that the requirement of intrathecal plain bupivacaine dose could be largely predicted by vertebral column length and abdominal girth for lower limb surgeries.
A previous study showed that the volume of lumbosacral cerebrospinal fluid was the primary determinant of sensory block level during spinal anesthesia [10], and there was broad variation in the volume of lumbosacral cerebrospinal fluid between individuals, which decreased following an increase in intra-abdominal pressure [11]. Greater abdominal girth was associated with a more notable increase in the intra-abdominal pressure [17]. Logic might suggest that there would be a smaller volume of lumbosacral cerebrospinal fluid in patients with shorter vertebral column lengths. Thus, to a great extent, abdominal girth and vertebral column length could accurately predict the dosage of intrathecal plain bupivacaine for T 12 and T 10 block level.
Many studies have investigated the effects of a patient's age [13], height [1,12], weight and BMI [1,12,14] on the spread of plain bupivacaine spinal anesthesia, but the results have been contradictory, or the variables have demonstrated to be weakly correlated with the spread of spinal anesthesia. Using three-dimensional magnetic resonance imaging, Sullivan and his colleagues [18] further proved that only weak correlations existed between the volume of lumbosacral cerebrospinal fluid and height, weight or BMI. In our previous and current studies, multiple linear regression analysis also showed that among the five predictors, age, weight and height had low value for predicting the dosage of intrathecal bupivacaine for the level of anticipant spinal anesthesia.
To date, some studies have been performed to investigate the optimal dose of intrathecal plain bupivacaine for certain surgeries [5,19]. However, these studies have only demonstrated the range of the intrathecal bupivacaine (e.g., ED 50 or ED 95 ) but not the optimal individual Table 2 The relationship of patient variables with the bupivacaine dose for loss of pinprick discrimination at T 12    dose. 5 Indeed, the ED 50 or ED 95 values could provide some guidance for using local anesthetics, but even if low doses of intrathecal plain bupivacaine were used, hypotension could still occurred because of excessive cephalad spread [19].
In previous studies, up-down sequential analysis [20] and logistic regression analyses [5] have been used to determine the optional dose of local anesthetics. Neither method could determine the optimal dose for an individual because the individual response to intrathecal local anesthesia varies greatly. In the current study, multiple linear regression analysis was performed to investigate the combined linear contribution of five patient variables to the volume of 0.5 % plain bupivacaine for loss of pinprick discrimination at T 12 and T 10 , then the regression equations between the 0.5 % plain bupivacaine volume and vertebral column length, abdominal girth were obtained. The R 2 for loss of pinprick discrimination at both T 12 and T 10 was approximately 0.9, indicating that the regression equation could accurately predict the dosage of specific intrathecal bupivacaine for T 12 or T 10 spinal spread in an individual.
In this study, reverse thoughts were used. We set T 10 spinal block level as the target level, and the bupivacaine dose administered was based on the experience of the anesthetist. No matter how much the bupivacaine dose were used, there were many possibilities that some patients , spinal upper block level would be T 12 or T 10 . When the sample size, based on spinal spread for loss of pinprick discrimination at T 12 and T 10 , reached 91 respectively could the study be terminated. Finally, 111 patients with loss of pinprick discrimination at T 12 and 121 patients with loss of pinprick discrimination at T 10 were enrolled in our study. 391 patients with loss of pinprick discrimination for other spinal levels were not analyzed.
Reynolds F. [21] advised avoiding spinal injection above L 3 to decrease the risk of spinal cord trauma. Therefore, the L 3 -L 4 interspace is usually the targeted intervertebral needle insertion site during spinal anesthesia, and it was adopted in our study. The accuracy of the regression equation for the dosage of intrathecal bupivacaine and vertebral column length and abdominal girth for T 10 and T 12 spinal spread levels should be confirmed in the future with larger samples. The subjects of this study were 19 to 60 years old; therefore, the accuracy of the regression equation for patients of other ages should also be confirmed.

Conclusion
In conclusion, our data indicated that vertebral column length and abdominal girth were strongly correlated with the dosage of intrathecal plain bupivacaine for the loss of pinprick discrimination at T 12 and T 10 . The two regression equations were Y T12 = 3.547 + 0.045X 1 -0.044X 2 and Y T10 = 3.848 + 0.047X 1 -0.046X 2 (Y, 0.5 % plain bupivacaine volume; X 1 , vertebral column length;and X 2 , abdominal girth), which can accurately predict the minimum and suitable intrathecal bupivacaine dose for lower limb surgeries to a great extent separately.