|
 
 |
| ORIGINAL ARTICLE |
|
| Year : 2017 | Volume
: 10
| Issue : 1 | Page : 84-90 |
|
Management of early postoperative hypoxemia: a comparative performance of Hudson face mask with nasal prongs
Suleiman A Adetunji1, Oyebola O Adekola2, Ibironke Desalu2, Olushola T Kushimo2
1 Department of Anaesthesia, Lagos University Teaching Hospital, Lagos, Nigeria
2 Department of Anaesthesia, Lagos University Teaching Hospital, Lagos; Department of Anaesthesia, College of Medicine, University of Lagos, Lagos, Nigeria
| Date of Web Publication | 3-Aug-2018 |
Correspondence Address:
Suleiman A Adetunji
Department of Anaesthesia, Lagos University Teaching Hospital, PMB - 12003, Surulere, Lagos
Nigeria
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/1687-7934.238469
Introduction Early postoperative hypoxemia may occur when patients breathe room air during their initial recovery period. Prolonged hypoxemia can result in delirium, dysrhythmia, and cardiac arrest.
Aim The aim of the present study was to compare the performance of face mask compared with that of the nasal prong in the management of early postoperative hypoxemia.
Patients and methods All procedures were performed using standard anesthetic and surgical techniques modified to the specific procedures. All patients had peripheral oxygen saturation (SpO2) at least 97% before being transferred to the recovery room. On arrival to the recovery room, 120 patients whose SpO2 decreased up to 94% were randomly allocated to either the face mask or nasal prong groups. They were commenced on oxygen therapy at 4 l/min through either device. A modified visual analogues scale was used to evaluate the level of comfort during oxygen therapy.
Results Early postoperative hypoxia occurred in 18.1%. The increase in oxygen saturation after commencement of oxygen therapy was significantly faster with nasal prongs (0.63±1.42 min) than with face mask (1.78±1.10 min) (P=0.001). The maximum SpO2 obtained was significantly higher with nasal prongs (98.77±1.29%) than with face mask (97.63±1.89%) (P<0.001). There was no significant association found between early postoperative hypoxemia and site or duration of surgery, as well as the volume of intravenous fluid (crystalloids) administered intraoperatively (P>0.05). Nasal prongs (91.7%) were significantly more comfortable compared with face mask (61.7%) (P=0.001). We have demonstrated that the use of nasal prongs was more efficient and comfortable compared with face mask in the management of early postoperative hypoxemia.
Keywords: early postoperative hypoxemia, face mask, nasal prongs
How to cite this article:
Adetunji SA, Adekola OO, Desalu I, Kushimo OT. Management of early postoperative hypoxemia: a comparative performance of Hudson face mask with nasal prongs. Ain-Shams J Anaesthesiol 2017;10:84-90 |
How to cite this URL:
Adetunji SA, Adekola OO, Desalu I, Kushimo OT. Management of early postoperative hypoxemia: a comparative performance of Hudson face mask with nasal prongs. Ain-Shams J Anaesthesiol [serial online] 2017 [cited 2023 Dec 5];10:84-90. Available from: http://www.asja.eg.net/text.asp?2017/10/1/84/238469 |
| Introduction | |  |
Early postoperative hypoxemia, which occurs within minutes to 2 h after surgery, has an incidence of 35–60% [1],[2],[3]. Hypoxemia may increase the risk for surgical wound infection, reduce anastomosis integrity, and result in poor wound healing [4],[5],[6]. It may also contribute to loss of gastrointestinal mucosal integrity, thus causing bacterial translocation into the circulation, leading to sepsis [4],[5],[6].
Hypoxemia may decrease cognitive function and cause delirium, which can cause patients to remove nasogastric tubes, surgical drains, and intravascular devices. Oxygen therapy has been reported to be beneficial in the treatment of postoperative delirium secondary to hypoxia [5],[7],[8]. Other secondary effects of hypoxemia include increased production of catecholamine, hypertension, tachycardia, cardiac arrhythmia, and myocardial ischemia [5],[7],[8]. In addition, there was an increase in both time of recovery room stay and the incidence of admission in ICUs [5].
Oxygen therapy must be administered at an appropriate dose and using an appropriate device for effective outcome [4],[5]. The modes of delivery of oxygen to spontaneously breathing, nonintubated patients during their presence in the recovery room include nasal cannula, nasal catheter, nasal prongs, face mask, or face tent [4],[5].
At our institution, variable performance oxygen devices such as Hudson face mask and nasal prongs are often used in the recovery room in patients at risk for postoperative hypoxemia, such as the elderly, sickler, and patients who underwent thoracic or upper abdominal surgeries. The choice of oxygen therapy usually depends on the experience of the anesthetist, availability of the device, and cooperation of the patient. The aim of this study was to compare the performance of Hudson face mask and nasal prongs in the management of postoperative hypoxia.
| Patients and methods | | |
This was a comparative study on the performance of Hudson face mask and nasal prongs in the management of early postoperative hypoxemia. The study was conducted at the Lagos University Teaching Hospital between October 2013 and October 2014. The approval of the Institutional Human Research and Ethics Committee and written informed consent were obtained.
Individuals of American Society of Anesthesiology (ASA) physical status I and II aged 18–80 years undergoing elective surgical procedures under general anesthesia using muscle relaxants were recruited. Those excluded from the study included patients with abnormal electrolyte/urea, upper respiratory tract infection, chronic cigarette smokers, chronic obstructive pulmonary disease patients, patients with sickle cell disease, preoperative hypoxemia, and sepsis, individuals of ASA III to IV status, those with anemia following hemorrhagic shock, as seen in ectopic pregnancy, and those who underwent post-thoracotomy and neurosurgery. In addition, those undergoing regional anesthesia and patients in whom the use of pulse oximeter may result in inappropriate results (peripheral vascular disease, severe anemia, or recent use of colored dyes) were excluded.
For the purpose of this study the following definitions were used:
Hypoxemia was graded into four values of peripheral oxygen saturation (SpO2): mild (86–90%), moderate (81–85%), severe (76–80%), and extreme (<80) [9].
The trigger value of oxygen saturation for commencement of oxygen therapy was SpO2 less than 94% [9].
The level of consciousness or sensorium was graded as follows: (a) fully awake; (b) asleep, but easily aroused; and (c) asleep and difficult to arouse [10].
The modified visual analogue scale (VAS) was used to assess the patient’s comfort in the use of the oxygen therapy devices (modified VAS score, 100 mm=worst possible discomfort, 0 mm=most comfortable).
Anesthetic technique
In the preoperative period, all patients were educated on the use of VAS to assess their level of comfort during the use of the oxygen therapy devices. Diazepam 5 mg oral was administered at night and on call to theater. On arrival to the theater, standard monitoring of the blood pressure (BP), heart rate (HR), and SpO2 was performed every 5 min with a multiparameter monitor (Datex Ohmeda Cardiocap 7100; Metropolitan Medical Services, Asheville, North Carolina, USA) until the end of surgical procedure.
Preoxygenation was carried out with oxygen 6 l/min for 3–5 min, followed by induction of general anesthesia with intravenous sodium thiopentone (5 mg/kg), and endotracheal intubation was facilitated with pancuronium (0.1 mg/kg). Anesthesia was maintained with isoflurane (1.5–2%) in 100% oxygen according to our institution protocol as there was no facility for nitrous oxide and air during the study period. Analgesia was induced through a multimodal approach, which included intravenous diclofenac 1 mg/kg, tramadol 1 mg/kg, and paracetamol 15 mg/kg intravenous over 15 min. Intravenous fluids were given based on maintenance of 110 ml/h and ongoing loses (from nasogastric tube and surgical drains). It is suggested that, in patients who received neuromuscular blockers, monitoring of reversal with a peripheral nerve stimulator should be encouraged; however, there was no such facility at the time of the investigation. At the end of surgery, residual neuromuscular blockade was reversed with intravenous atropine 0.02 mg/kg and neostigmine 0.04 mg/kg. After extubation, all patients were given 100% oxygen using a well-fitted face mask through the anesthetic machine for 3 min. Oxygen saturation was recorded at the end of the 3 min (time 0). All patients were ensured to have maintained SpO2 at 97% or more before being transferred to the recovery room by the researcher. During transfer, the patients were continuously monitored for SpO2 using a pulse oximeter (Nonin Avant 9600; Medical Product Service GmbH, Borngasse, Germany). The alarm on the pulse oximeter was preset at SpO2 up to 95%, and thus when the SpO2 fell below 95%, it was recorded. The frequency of hypoxemic events during transportation of patients was noted from the monitor and recorded. Supplemental oxygen was not administered during transport.
On arrival to the recovery room, standard monitoring of SpO2, BP, HR, and respiratory rate (RR) was carried out. The level of consciousness on arrival using postanesthetic recovery score was also monitored [11]. Patients with oxygen saturation at least 94% underwent continuous monitoring as per standard protocol in the recovery room. However, those with SpO2 less than 94% underwent humidified supplemental oxygen therapy commenced using variable performance oxygen devices, face mask or nasal prongs, at an oxygen flow rate of 4 l/min. The patients were randomly assigned to two groups by the researcher, from a sealed envelope, which contained the groups written on a folded piece of paper. Group A (face mask) received oxygen therapy through Hudson face mask, and group B (nasal prongs) received oxygen therapy through nasal prongs. Continuous monitoring of vital signs and the sensorium was continued every 5 min in the recovery room for 30 min. Thereafter, patients with SpO2 greater than 94% were monitored every 15 min until they were discharged from the recovery room.
The efficacy of the devices was determined by comparing the proportion of the participants who desaturated while on oxygen therapy, the time required to achieve a saturation at least 95%, and the ability of the participant to maintain SpO2 greater than 94% without carbon dioxide retention. The ability to provide good comfort/convenience for the patient during the procedure was assessed by the use of a modified VAS for 2 h in the recovery room before their discharge to the ward [12].
The data collated included demographic data such as age, weight, height, BMI, and hemoglobin concentration. Other data included the type, duration, and site of surgery and incision, the presence of shivering, the choice of oxygen therapy device, the presence of carbon dioxide retention, and the type and volume of intraoperative fluids transfused.
Statistical analysis
The sample size calculation was based on a predetermined statistical formula used for the comparison of proportion of two independent groups [13]. The reported incidence of early postoperative hypoxemia was 35–60% [2],[3],[4], with an average incidence of 45% (π1=0.45). We aimed to achieve a reduction of 20% incidence (π2), using α=95% and β=80%; a sample size of 60 participants per group was considered appropriate for the study.
Data were expressed as means±SD, frequency, and percentile as indicated. Student’s t-test was used for comparison of means, while whereas the χ2-test was used to compare frequencies. The risk for development of hypoxia, such as the site of surgery, duration of surgery, volume of fluid used intraoperatively, and the type of incision made, were analyzed using univariate analysis. A P value of less than 0.05 was considered significant for all tests. All analyses were performed using the statistical package for the social sciences (SPSS) for Windows version 17 (SPSS Inc., Chicago, Illinois, USA).
| Results | | |
A total of 120 adult patients were enrolled in this study, including 60 adult patients for each group throughout the course of the study. The mean age, weight, BMI, and preoperative hemoglobin concentration were comparable between the groups. Similarly, female preponderance was noted (24 : 36 in group A; 14 : 46 in group B) in both groups ([Table 1]). The most common surgery was pelvic surgery (myomectomy) in both groups (62 cases, 51.67%), followed by mastectomy (14 cases, 11.67%), varicocelectomy (18 cases, 15%), and others (tonsillectomy, laryngoscopy, biopsy, and excision of lipoma) (24 cases, 20%), and the least common surgery was thyroidectomy (two cases, 1.66%). | Table 1 Demographic data and clinical characteristics of the face mask group versus the nasal prong group
Click here to view |
The oxygen saturation was observed to be highest (99–100%) at the immediate postextubation period following 3 min on supplemental oxygen (time 0). Thereafter, saturation declined steadily during transportation ([Figure 1]). | Figure 1 Mean oxygen saturation trend during theater/recovery room transit. Time 0 corresponds to oxygen saturation after extubation and 3 min of supplemental oxygen after extubation
Click here to view |
[Figure 2] shows the comparative performance of the two oxygen therapy devices during oxygen therapy in the recovery room. Although the mean oxygen saturation decreased before the administration of oxygen (preoxygen therapy) in both groups, the difference was nonsignificant between the face mask group (93.2±7.0%) and the nasal prong group (93.3±2.6%) (P=0.461). However, the patients in the nasal prong group responded faster (0.63±1.42 min) to oxygen therapy than those in the face mask group (1.78±1.10 min) (P=0.001). A significantly higher SpO2 value (97.4±2.5%) was achieved in the nasal prong group than in the face mask group (96.0±3.5%) at 5-min duration of oxygen therapy (P=0.002). There was no incidence of hypoxemia or desaturation when patients were on oxygen therapy with either of the devices during the study ([Table 2]). | Figure 2 Comparison of peripheral oxygen saturation (SpO2) during oxygen therapy with the two devices. Statistically significant difference at 5 min (P=0.002)
Click here to view |
 | Table 2 The comparison of the efficacy of SpO2 for the face mask group versus the nasal prong group
Click here to view |
The mean HR was significantly higher in the face mask group than in the nasal prong group at 5, 10, 30, and 45 min in the recovery room ([Figure 3]). There was a gradual drop in mean arterial blood pressure (MAP) during the study period. Although patients in the face mask group exhibited higher mean values, this was not statistically significant until 30 min into the study when the mean MAP was significantly higher in the face mask group (70.4±16.5 mmHg) than in the nasal prong group (57.9±16.2 mmHg) (P=0.008) and also at 45 min (69.8±15.5 mmHg and 51.1±30.0 mmHg, respectively; P=0.001) ([Table 3]). | Figure 3 Comparison of mean heart rate with the two devices. Values represent mean±SD heart rate and P value for both devices
Click here to view |
 | Table 3 The comparison of the mean arterial blood pressure between the face mask group and the nasal prong group
Click here to view |
The mean RR was significantly higher in the face mask group than in the nasal prong group at 5, 25, 30, 45, and 60 min (P≤0.05) ([Table 4]). | Table 4 The comparison of mean respiratory rate during oxygen therapy between the face mask group and the nasal prong group
Click here to view |
Using univariate analysis to determine the risk factors for the development of postoperative hypoxemia, none of the confounding variables investigated were implicated ([Table 5]). | Table 5 The univariate analysis of risk factors in the development of postoperative hypoxemia
Click here to view |
The mean level of convenience using modified VAS was similar in the face mask group (0) versus the nasal prongs group (0). However, a higher proportion of patients in the nasal prong group (55 cases, 91.7%) were comfortable with the device compared with the face mask group (37 cases, 61.7%) (P=0.001) ([Table 6]). | Table 6 Comparison of patient satisfaction between the face mask group and the nasal prong group
Click here to view |
The nasal prongs cost ≈6 USD per unit, whereas the facemask costs ≈7.5 USD per unit.
| Discussion | | |
We have demonstrated that early postoperative hypoxia occurred in 18.07% of patients, which is within the stated range of 7.8–60% [1],[2],[3]. The wide variation in incidence has been attributed to differences in participant cohort, type of oxygen delivery device, and anesthetic and surgical techniques. The lower incidence of hypoxemia in the present study was attributed to the maintenance of anesthesia with isoflurane in 100% oxygen. The observation might differ with the use of nitrous oxide for maintenance, because the latter is associated with diffusion hypoxia if oxygen (100%) is not administered for about 10 min after its discontinuation. Other factors that may increase the incidence of early hypoxemia include residual effect of incomplete reversal of neuromuscular agent. In the present study, there was no patient who had oxygen saturation up to 90% during the theater–recovery room transit period; moreover, no patient desaturated up to 94% (our oxygen therapy trigger). Thus, oxygen therapy may not be required during transit to the recovery room postoperatively. This has previously been reported by other researchers in the subregion [3]. The use of maneuvers such as placing patients in a recovery position or propping-up during transport, which prevents airway obstruction, and splinting of the diaphragm may just be enough to prevent hypoxemia or desaturation, which may occur during transport. Nevertheless, it has been reported that when the duration of transport without supplemental oxygen is prolonged, the risk of developing postoperative hypoxemia is increased [3].
Other factors implicated in the occurrence of desaturation or hypoxemia during transport include premorbid state of the patient (preoperative hypoxemia and airway obstruction), hypoventilation secondary to the residual effect of inhalation agents, and the use of opioids and neuromuscular blockers. To reduce the influence of confounding factors, only ASA I and II patients were recruited. It was observed that desaturation in the recovery room occurred within 30 min of arrival. This is similar to a previous observation that early postoperative hypoxemia occurred most commonly less than 1 h after anesthesia in older children and adults [2]. During oxygen therapy with either of the devices, none of our participants developed hypoxemia. This may suggest that either the nasal prongs or face mask were effective in delivering oxygen in the early postoperative period in patients with oxygen saturation up to 94%. The findings in the present study are in agreement with previous observations [2],[10],[14]. However, a high incidence of hypoxemia (25%) was observed with the use of the aerosol face tent during oxygen therapy [15]. This may be because the design of the face tent allows significant air entrainment, thereby reducing the fractional inspired oxygen concentration (FiO2) [14].
The nasal prong was observed to be more efficient compared with the face mask for oxygen therapy in the present study. This could be due to a larger dead space volume reservoir of the face mask, which takes longer to be filled, compared with the smaller nasopharyngeal reservoir of the nasal prong. It is also possible that the claustrophobic effect of the face mask affected patient compliance and hence oxygen delivery (FiO2) to the patients.
It has been documented that arterial blood gas (ABG) analysis of oxygen pressure (PaO2) is the gold standard for the evaluation of oxygen therapy [12],[16]. Nevertheless, several scholars have reported that the results obtained from pulse oximeter (SpO2) correlated well with ABG PaO2 [12],[16].
In the present study it was observed that there was no association between early postoperative hypoxemia and sex, volume of intraoperative infusion, and duration and site of surgery. This may be because surgeries involving big cavity were excluded, as a high proportion (>50%) in the present study underwent myomectomy and exploratory laparotomy. A similar observation was reported by other scholars in relation to the duration and site of surgery [3],[9]. In contrast, the site of surgery has been implicated in the development of hypoxemia in the late postoperative periods [5],[16]. It has been reported that early postoperative hypoxemia is mainly due to anesthetic factors, whereas late postoperative hypoxemia is related to reduced functional residual capacity of the lung, especially with upper abdominal surgeries [5].
The ability of the devices to cause carbon dioxide retention (the incidence of development of carbon dioxide retention in both devices) was evaluated with changes in HR and BP in the patients during oxygen therapy. In the present study, a significant increase was observed in the HR, MAP, and RR within the first 5 min of commencement of oxygen therapy with the two devices. The increase was higher in the face mask group compared with the nasal prong group. These initial cardiovascular changes during oxygen therapy could be attributed to the period in which the reservoir (dead space) of the device was equilibrating with oxygen, which was more pronounced in the face mask group. This observation was corroborated by early researchers such as Bethune and Collins [17], who noted that some variable performance devices can cause rebreathing (carbon dioxide retention). However, there is paucity of knowledge on how best to measure volume of rebreathing through the direct measurement of end-tidal carbon dioxide while patients are on these devices. Waldau et al. [18] made an attempt using oxygraph, a modern gas analyzer, to measure the end-tidal oxygen fraction (FeCO2) and then calculate the end-tidal carbon dioxide (FeCO2) from alveolar gas equation. However, their results were based on mathematical assumptions, which may not hold good in clinical situations in which these devices were being used. It is also possible that the increase in HR and the BP noted with the use of face mask above is due to an increase in the work of breathing. Jensen et al. [19] have demonstrated in healthy volunteers that minute ventilation is greater when using a face mask at a flow rate less than 5 l/min compared with that when no face mask was used. They, however, observed no changes in PaCO2, measured using ABG analysis.
The use of HR and BP to monitor rebreathing in the present study may not be ideal, as other clinical conditions such as pain can result in such changes. The ABG analysis would have been more ideal to measure PaCO2 to confirm carbon dioxide retention.
The increase in HR, BP, and RR observed during oxygen therapy may be hazardous in patients with poor cardiopulmonary reserve and coexisting medical conditions such as hypertension and diabetes mellitus. A previous study has reported an association between the presence of perioperative tachycardia and the development of perioperative myocardial ischemia [5].
The results in the present study showed that the use of nasal prongs was associated with better satisfaction for our patients compared with the face mask; this result is similar to the findings by other researchers [2],[10]. Nasal prongs have also been shown to be cost-effective, when compared with the face mask. This is very important in this period of global economic recession.
This study is limited by the fact that the assessment of adequacy of oxygenation was determined with the pulse oximeter. In the present study the arterial blood analyzer was not used due to the cost implication and invasiveness of the technique, when its use was not indicated. The clinical signs of carbon dioxide retention were used instead of a capnography or capnometer, which were unavailable in our recovery room at the time of our study. Despite this, we have demonstrated that the nasal prongs were a more effective device of delivering oxygen in patients with early postoperative hypoxia; it is also comfortable and cost-effective.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Sia S, D’Andrea V, Mamone D, Pagnotta L, Verre M. Early postoperative hypoxemia: incidence and effectiveness of oxygen administration. Minerva Anestesiol 1994; 60:657–662.  |
| 2. | Xue FS, Huang YG, Tong MY, Liu QH, Liao X, An G et al. A comparative study of early postoperative hypoxemia in infants, children, and adults undergoing elective plastic surgery. Anesth Analg 1996; 83:709–715. |
| 3. | Amanor-Boadu S, Osinaike B, Oyeleke S. Routine oxygen administration in the PACU: should everybody have it? Int J Third World Med 2006; 3:1–6. |
| 4. | Strachan L, Noble DW. Hypoxia and surgical patients − prevention and treatment of an unnecessary cause of morbidity and mortality. J R Coll Surg Edinb 2001; 46:297–302. |
| 5. | Ayhan H, Iyigun E, Tastan S, Orhan ME. Comparison of two different oxygen delivery methods in the early postoperative period: randomized trial. J Adv Nurs 2009; 65:1237–1247. |
| 6. | Waldau T, Larsen VH, Bonde J. Evaluation of five oxygen delivery devices in spontaneously breathing subjects by oxygraphy. Anaesthesia 1998; 53:256–263. |
| 7. | Deiner S, Silverstein J. Postoperative delirium and cognitive dysfunction. Br J Anaesth 2009; 103(Suppl 1):i41–i46. |
| 8. | Gift AG, Stanik J, Karpenick J, Whitmore K, Bolgiano CS. Oxygen saturation in postoperative period; patients at low risk for hypoxemia: is oxygen therapy needed? Anesth Analg 1995; 80:368–372. |
| 9. | Kitamura H, Sawa T, Ikezono E. Postoperative hypoxemia: the contribution of age to maldistribution of ventilation. Anaesthesiology 1972; 36:244–252. |
| 10. | Aakerland LP, Rosenberg J. Postoperative delirium: treatment with supplementary oxygen. Br J Anaesth 1994; 72:286–290. |
| 11. | Gennari R, Alexander JW. Effects of hyperoxia on bacterial translocation and mortality during gut-derived sepsis. Arch Surg 1996; 131:57–62. |
| 12. | Razi E, Akbari HA. Comparison of arterial oxygen saturation measured both by pulse oximeter and arterial blood gas analyzer in hypoxaemic and non-hypoxaemic pulmonary disease. Turk R J 2006; 7:43–47. |
| 13. | Allan D. Approaches to sample size estimation in design of clinical trial − a review. Stat Med 1984; 3:199–214. |
| 14. | Russell GB, Graybeal JM. Hypoxemic episodes of patients in the postanesthesia care unit. Chest 1993; 104:899–903. |
| 15. | Stausholm K, Rosenberg-Adamsen S, Skriver M, Kehlet H, Rosenberg J. Comparison of three devices for oxygen administration in the late postoperative period. Br J Anaesth 1995; 74:607–609. |
| 16. | Akinyemi OA, Sanusi AA, Eyelade OR. Evaluation of a modified APACHE II scoring system in the intensive care unit of a tertiary hospital in Nigeria. Afr J Anaesth Intens Care 2011; 11:1–6. |
| 17. | Bethune DN, Collins JM. Comparative performance of six oxygen delivery equipments. Thorax 1967; 22:221. |
| 18. | Waldau T, Larsen VH, Bond J. Evaluation of five oxygen delivery devices in spontaneously breathing subjects by oxygraphy. Anaesthesia 1998; 53:256–263. |
| 19. | Jensen AG, Johnson A, Sandstedt S. Rebreathing during oxygen therapy with face mask. The effect of oxygen flow rates on ventilation. Acta Anaesthesiol Scand 1991; 35:289–292. |
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]
|
Search Pubmed for