Esibov A et al. Resuscitation 2015;91:116-121.

• The LINC trial was designed to compare manual CPR and mechanical CPR. Adult patients with unexpected out-of-hospital cardiac arrest were started on manual CPR and immediately randomised into either group using sealed opaque envelopes at the patient’s side.
• Patients with traumatic cardiac arrest, pregnancy, extreme body habitus, defibrillation before the device arrived on scene or crew-witnessed cardiac arrest were excluded from the study.
• Primary out was defined as 4-hour survival after successful ROSC. Secondary outcome included ROSC, arrival to emergency department with a pulse, survival with good neurological recovery. In terms of both primary and secondary outcomes, there was no significant difference between the two groups.
• A point to note from the LINC trial is that the manual CPR and mechanical CPR groups utilise two different protocols. The protocol for manual CPR group is similar to the 2005 European Resuscitation Council guidelines. The protocol for mechanical CPR included 3 minutes of mechanical compression, first defibrillation after 90s during ongoing mechanical compressions regardless of initial rhythm, first rhythm check after 3 minutes of mechanical compression.
• Link to article

Leung AA et al. Journal of Pain Research 2018;11:2477-2489.

• This meta-analysis was conducted to investigate the effectiveness of TAP block after hysterectomy:
• Patients – Initial search from PubMed (n=364), MEDLINE (n=53), EMBASE (n=95), Cochrane Library (n=73), Web of Science (n=49), and ClinicalTrials.gov (n=7) databases; 13 prospective randomized controlled studies published from 2008 to 2016; 841 participants were included. Types of hysterectomy: 8 studies underwent total/partial AH and 5 studies underwent LH/RH. All used standard GA
• Intervention - Pre-/post-op TAP block
• Control – No or sham block
• Outcome (Primary) Mean 24-hr morphine consumption
• Secondary outcomes: Time to first request for analgesic, Pain score at rest, Pain score on movement at different times postoperatively, including 2, 4, 6–8, 12, 24, and 48 hours. Quality of recovery score (QoR-40) –> 40-item questionnaire intended to measure quality of recovery grouped according to various aspects (dimensions) of recovery: i) emotional state (n 9) ii) physical comfort (n 12) iii) psychological support (n 7) iv) physical independence (n 5) v) pain (n 7). Adverse events: Rate of N&V
• Search for relevant studies comprehensive?
  • Initial search from PubMed (n=364), MEDLINE (n=53), EMBASE (n=95), Cochrane Library (n=73), Web of Science (n=49), and ClinicalTrials.gov (n= 7) databases
  • Search results were limited to English
  • Two authors (HZ and JP) performed the searches and independently reviewed and extracted data from each study
  • Potential sources of methodological bias according to the Cochrane Collaboration Risk of Bias tool were evaluated in all the included studies
• Were the inclusion criteria used to select articles appropriate?
  • Inclusion criteria
    Including women who underwent hysterectomy regardless of abdominal, laparoscopic, or robotic-assisted surgery
    Including women who received TAP block pre- or postoperation
    Being an RCT
    Being a comparative study between TAP block and sham or no block groups
  • Exclusion criteria
    Being a non-randomized study
    Failing to focus on TAP block compared with sham block or no block
    Being a case report, review paper, or unpublished literature
    Undergoing a surgery other than hysterectomy
• Were the included studies of good quality?
  • All are RCTs. Overall, the majority of trials had a low risk of bias
• Were the results similar from study to study?
  • No; because of methodological heterogeneity.
  • Chi square test: Assume null hypothesis that all studies are homogenous/each study is measuring an identical effect –> If p-value <0.1 –>reject null hypothesis –>significant heterogeneity
  • I2 statistics: < 25% –> low heterogeneity
• Conclusion
  • TAP block can provide effective analgesia for AH
  • BUT the effect of TAP block in LH/RH is limited
  • TAP block significantly reduced mean 24-hour morphine consumption and prolonged the time to first request for analgesic in AH only, and TAP block has limited effectiveness for participants who underwent LH or RH
  • TAP block also significantly reduced the pain scores at rest and on movement postoperatively at different times in AH but not in LH or RH
  • TAP block significantly reduced the rate of nausea and vomiting after the operation
• External validity: this meta-analysis had several limitations:
  • High heterogeneity – factors such as different doses and types of anesthetics and different postoperative analgesic methods may impact the heterogeneity of the meta-analysis
  • More RCTs with larger sample sizes in these fields are still needed to strengthen the results of meta-analysis
• Link to article

Leung AA et al. JAMA 2012;172:1474-1481.

• This study investigated whether preoperative mild (130-140mmol/L) or moderate-to-severe (<130 mmol/L) hyponatremia was associated with perioperative outcomes:
• Patients – the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) Participant Use Data Files-> Adults with preoperative sodium level (<135) undergoing surgery from 1/1/2005 to 31/12/2010
• Intervention - N/A
• Control – patients with preoperative normal sodium level
• Outcome (Primary) 30-day mortality; (Secondary) stroke, coronary events, pneumonia, wound infection, length of stay
• Study design:
  • Observational study – Retrospective cohort study in which the study is carried out at present time and look at the past for the data
  • Pros: examine multiple outcome, less costly than prospective cohort, study can be carried out in shorter time, immediate access to data
  • Cons: Null hypothesis –> there is no relationship between preoperative hyponatremia with post operative mortality
  • In this study 75423 patients with preoperative hyponatremia were compared with 888840 controls
• Statistical methods:
  • Multivariate logistic regression: to find an equation that best predicts the probability of y variable as the function of x variable
  • Non parametric quantile regression model : for studying length of stay outcome (non dichotomous data)
  • Subgroup analysis to study whether baseline characteristics of the population group will affect the study result (ANA, BMI, health status , surgery type)
  • Sensitivity analysis for potential selection bias – the study was repeated twice with assuming the patients who were excluded from the initial cohort because of not having preoperative sodium level with (1) normal sodium level and (2) hyponatremia.
  • Missing values have undergone dummy coding
  • Secondary analysis for the primary outcome according to the number of days from the preoperative sodium measurement to the time of operation
• Results:
  • Primary outcome – 30day mortality: with mild hyponatremia (4.6% deaths; aOR, 1.38; 95% CI, 1.32-1.45); with moderate to severe hyponatremia (9.6% deaths; aOR, 1.72; 95% CI, 1.58-1.88) (P<0.001)
  • Secondary outcomes – Coronary events, wound infection, pneumonia with aOR >1 and 95% CI not crossing 1 in preoperative hyponatremia patients. Hypo-natremia was typically associated with longer median lengths of stay by approximately 1 day for most surgical procedures during the study
• Definition of "Odds Ratio":
  • Odds Ratio = Odds of exposure in cases / odds of exposure in control
  • In this study, preoperative hyponatremia is associated with increased 30-day mortality even after subgroup analysis.
• Pros of the study:
  • Cohort
  • Large sample size
  • Good subgroup analysis to study baseline characteristics of population group
  • Investigate the effect of outcome with different degrees of hyponatremia (mild and moderate to severe group)
  • Sensitivity analysis to minimize potential selection bias
  • Secondary analysis to study the primary outcome according to the number of days from the preoperative sodium measurement to the time of OT
• Cons of the study:
  • Not knowing the reason of hyponatremia
  • Not mentioning any treatment given for treating hyponatremia before surgery
  • Postoperative sodium level not known
  • Not representing a casual relationship between sodium level and perioperative mortally/morbidity
  • Individual may contribute to the study more than once
  • Anesthesia mode not mentioned (GA/RA)
• External validity
  • Study was performed in USA, mainly non Asian population
  • However as large sample size, therefore may still be applicable to our practices
  • Although this study didn’t tell whether treating hyponatremia preoperatively can improve the post-operative outcomes, preoperative sodium level may be a marker for poor prognostic outcome. There is no consensus for the preoperative sodium level, therefore anesthetists should assess the situation in a case-to-case basis, including balancing the risks and benefits of postponing the surgery for investigation and treatment of hyponatremia. But generally if elective surgery with moderate to severe hyponatremia <130 (or symptomatic hyponatremia), then should consider postponing surgery for further investigations and management
• Link to article

Lin MH et al. Stroke. 2019;50:1364-1371.

• New onset perioperative atrial fibrillation (AF) might be a risk factor for stroke and mortality. This meta-analysis evaluates the relationship of perioperative AF with early and late risks of mortality and stroke. In this meta-anlysis, which included 35 studies with 2458010 patients, the results showed that POAF was associated with increased risks of early stroke (odds ratio, 1.62; 95% CI, 1.47–1.80), early mortality (odds ratios, 1.44; 95% CI, 1.11–1.88), long-term stroke (HR, 1.37; 95% CI, 1.07–1.77), and long-term mortality (HR, 1.37; 95% CI, 1.27–1.49).
• Internal Validity:
  • Patients – those with POAF
  • Intervention / Comparison – without POAF
  • Outcome (Primary) – early and late stroke risk/mortality
• Literature search: Pubmed, EMBASE, and Cochrane Library (1966 through March 2018) with no language restriction. Two investigators independently conducted the literature search. Any discrepant judgments were resolved through joint discussion. Studies found were then assessed for their quality by using the US preventive Task Force modified checklist.
• Heterogeneity: were the results similar from study to study? Look at forest plot
  • Graphical display of results of meta-analysis which allow visual comparisons of all the studies included
  • Individual studies are plotted as boxes with their size proportional to its weighing (sample size)
  • 95% Confidence intervals were plotted with vertical reference line representing "line of no effect". P value is calculated for the strength of overall effect
  • A diamond represents the 95% CI from the pooled results.
  • Heterogeneity calculated using tests e.g. Chi-square test, I-square test, Tau-square test
• If the heterogeneity tests are significant (e.g. in this study, I2 >75% in early mortality, long term stroke as well as long term mortality) means that the journals included are HETEROGENOUS, which is NOT GOOD.
• External Validity - several issues to consider:
  • Unknown diagnostic criteria on e.g. duration, episode, timing with reference to operation for "POAF"
  • How was the AF being treated?
  • Was underlying AF being picked up? (effect of pre-operative unknown chronic pAF instead of new onset perioperative AF)
  • 28 out of 35 studies included were on patients with underlying cardiac pathology undergoing CABG. Applicability to other patient groups undergoing other operations? Cardiac surgery and POAF may have be more related to peri-operative cardiac tissue inflammation, etc, which could go away post-operatively. But for non-cardiac surgeries, POAF may signify a more chronic persistent poor premorbid state and hence more long term effects.
• Implications to our clinical practice:
• Any peri-operative arrhthymia is to be treated due to high benefit to risks ratio. This study, due to its heterogeneity and therefore doubtful conclusion, as well as the poor applicability to our usual patient groups, probably fails to persuade a change in our usual clinical management, which would be to treat new-onset POAF by initially trying early medical or electric conversion, and if POAF recurs or is persistent, to treat with rate control and anticoagulation. The study does however triggers thoughts on the risks and significance of peri-operative arrythmia, and suggests that even if the arrythmia goes away, maybe we should still consider consulting medical for follow up management.
• Link to article

McCann ME et al. Lancet. 2019;393:664-677.

• Exposure to general anaesthetics leads to neurotoxicity in laboratory animals. However, it is controversial whether general anaesthesia in early infancy affects neurodevelopmental outcomes.
• The GAS Trial is an international randomized, controlled trial to compare the neurodevelopmental outcome at 5 years of age in patients who underwent short GA in early infancy for hernia repair VS awake-regional anaesthesia.
• The outcome studied, IQ score at 5 years, is a meaningful outcome as it has implication for long term intellectual function.
• Internal Validity:
  • Patients – 722 recruited; Infants<60 weeks’ postmenstrual age; born at >26 weeks gestation, planned inguinal herniorrhaphy, without previous exposure to GA or risk factors for neurological injury.
  • Intervention / Comparison – Awake regional anaesthetic vs Sevoflurance based general anaesthetic
  • Outcome (Primary) – Full-scale IQ on WPPSI-III at 5 years of age
• Null hypothesis: no clinically important difference in neurodevelopmental outcome between GA and RA
• Randomization: computer based stratified＋ block randomzation
  • 363 awake regional; 359 GA
  • Blocks of two or four in a computer-generated random-allocation sequence; allocation concealment
  • stratification by site and by gestational age at birth
• The groups were similar at baseline, and the groups were treated equally aside from the allocated treatment.
• 91 families were lost to follow-up in the awake-regional anaesthesia group and 97 in the general anaesthesia group (74% follow-up). Multiple imputation under multivariate distribution to impute the missing outcome data.
• 76 protocol violations: Most violation from RA group received GA due to inadequate/failed regional block Both per-protocol and intention to treat analysis performed.
• Blinding: Partial
  • Anaesthetist – aware of allocation (for sure)
  • Observer (administer neurodevelopmental assessment) blinded
  • Parents: informed if they asked about the group allocation (reasonable to know… What mode did the child receive), told to mask assessors about the allocation
• Results / Conclusions: Primary outcome was WPPSI-III FSIQ. Adjusted mean difference for awake-regional anaesthesia minus general anaesthesia 0·23, 95% CI –2·59 to 3·06. Intention-to-treat analysis (0·16,–2·45 to 2·78). Equivalence in WPPSI-III FSIQ at 5 years between children who received awake regional anaesthesia and GA for inguinal herniotomy (of slightly less than an hour) in infancy.
• External Validity
  • Is IQ score at 5 years meaningful? It is explained in the paper that IQ at age 5-6 strongly correlated with adult IQ – implication for effect on long term intellectual function
  • Is the study population similar to our patients? Male predominant study patients – explained by the higher prevalence of hernia in male. Early infancy – results cannot be translated to older children. Short operation (average OT time less than 1 hour) – cannot translate to longer operations.
  • Does the result of GAS trial correlate with other trials? e.g. MASK – Also no difference between child with single exposure vs no exposure; children that had multiple anaesthetics did have an increased risk of deficits in processing speed and fine motor outcomes, and parents reported increased problems related to executive function, behaviour, and reading; however the GAS trial was not designed to address the above issue, secondary outcomes did not showed increase in other behavioural/developmental deficits. PANDA – correlate with the finding of no difference in IQ in child who received exposue to GA to their unexposed siblings.
  • Implication to clinical practice: Counselling to parents: strong evidence that single episode, short duration (1 hour) GA in early infancy not having effect on child’s neurodevelopment in terms of intellectual function. Should not defer operation for the sole purpose of preventing such neurodevelopment effect – risk for delaying peration e.g. hernia complications. Still pending further research on effect on other neurocognitive domains, and the age/exposure threshold of such effect.
• Link to article

Baekgaard et al. Anesthesia & Analgesia. 2019;128:504-510.

• Systematic review: A review in which evidence on a topic or research question has been systematically identified, appraised and summarised according to pre-determined criteria. A systematic review may incorporate meta-analysis, but don’t have to.
• Meta-analysis: A statistical technique which combines the results of several studies into a single estimate, giving more weight to larger studies. Forest plots are commonly used to represent the meta-analysis of results
• This systematic review addressed the following:
  • Population – Adult trauma patients intubated in pre-hospital setting or in AED within 1h of arrival
  • Intervention – use of either ketamine
  • Comparison – vs any other induction agent
  • Outcome – Primary (30-day mortality) and Secondary (in-hospital mortality, info on blood transfusion, hospital stay length)
• There was a comprehensive literature search. Major bibliographic databases were searched (Cochrane, EMBASE, Pubmed), with no language restrictions. Search was done by 2 authors and interrater reliability was checked with Cohen kappa statistics.
• Selection criteria were appropriate. However, the number of studies included was limited - 3 cohorts and 1 RCT. The RCT in 2009 has low risk of bias; the Vietnam war cohort from 1976 has serious risk of bias; the other 2 cohorts have moderate risks of bias. Cochrane tool for assessing risk of bias was used for RCT's and Newcastle-Ottawa Scale for risk of bias was used for non-randomized trials. Because of methodological heterogeneity in terms of risk of bias in the studies it was not found meaningful to perform a meta-analysis on the studies.
• Conclusion: Existing evidence cannot exclude benefit or harm of ketamine compared to other drugs. Larger RCTs are warranted to establish more solid evidence.
• In terms of external validity, we thought that maybe the outcomes investigated in this study are not as interesting to anaesthetists (we may be more interested in intubation success/attempts/difficulties/desaturation episodes), but the outcomes in this study are admittedly more “significant” in terms of patient outcomes. Also, perhaps ketamine is not as widely utilised in AED/out of hospital setting and hence the question asked was not addressed in most literature.
• Link to article

Wijeysundera et al. Lancet 2018;391:2631-40.

• A multicentre, international prospective cohort study to compare preoperative subjective assessment with alternative markers of fitness, including Duke Activity Status Index (DASI) questionnaire, for predicting death or complications after major elective non-cardiac surgery.
• The study concluded that subjectively assessed functional capacity should not be used for preoperative risk evaluation. Subjective assessment had 19.2% sensitivity and 94.7% specificity for identifying inability to attain 4METS during cardiopulmonary exercise testing. Only DASI scores were associated with predicting the primary outcome. Hence, the authors also concluded that clinicians could use a measure such as DASI for more accurate cardiac risk assessment.
• Advantages of cohort study: finding the risk factors; temporal relationship between risk factors & outcomes; sometimes RCT is not ethical to perform; able to study rare exposure; able to measure multiple outcomes
• Disadvantages of cohort study: need large sample size; need long follow-up duration; not good for disease with long latency; potential lost to FU; high study cost; susceptible to selection bias
• Link to Duke Activity Status Index questionnaire: Here
• Link to tool for critical appraisal of cohort studies: Here
• Link to article

Birenbaum et al. JAMA Surgery 2019;154(1):9–17.

• In this large non-inferiority RCT comparing cricoid vs sham procedure on the rate of pulmonary aspiration, the study investigators failed to demonstrate the non-inferiority of the sham procedure in preventing pulmonary aspiration.
• Non-inferiority trial design, particularly its applications and statistical background including the slightly unusual null/alternative hypotheses were discussed.
• Factors affecting sample size calculation: effect size, alpha (predefined significance level), beta (lower beta error means higher power), variance, test sensitivity and design, as well as non-statistical factors. The actual pulmonary aspiration rate from the study data was lower than that predicted by the investigators, so the sample size was inadequate.
• Intention-to-treat ITT vs Per-protocol PP analysis: ITT is usually the preferred less biased analytical approach for superiority trials. (i.e. analyse all randomized patients regardless of whether they received the assigned intervention or not). However for non-inferiority trials, ITT which includes dropouts etc. tends to bias the results towards equivalence. A PP analysis which analyses patients who had no major protocol violations is more likely to identify any true treatment differences, although there could be bias in either direction. Therefore, it is recommended for non-inferiority trials to perform both analyses and draw conclusion if both produce the same results.
• For the external validity, it is reasonable to conclude that cricoid pressure is still a safe option for patients with aspiration risk, since from the secondary outcomes the median intubation time for the cricoid group was only longer by 4s and rates of difficult intubation or impossible intubation were not different between groups.
• Link to article

Forster C et al. British Journal of Anaesthesia 2018 Nov 1;121(5):1059-64.

• In terms of internal validity, Student’s T-test may not be appropriate to analyse the primary outcome data as the population may not be normally distributed
• Significant reduction of propofol dose in the study may not translate to meaningful clinical outcome; the more meaningful question may be the hemodynamic or respiratory effects of decreasing the consumption of propofol
• Link to article

Myles PS et al. New England Journal of Medicine 2018 May 10.

• A landmark study comparing restrictive vs liberal therapy for major abdominal surgery in at-risk patient groups
• Randomisation – highlighted the differences between simple randomisation, block randomisation and stratified randomisation
• Blinding was not possible in this study
• There is inadequate explanation from the authors about maintaining haemodynamic targets within the usual practical/ethical standards, for example, in the restrictive group the investigators restricted fluid regardless of the hemodynamic or other clinical parameters such as oliguria — and treated with pressors instead. It may seem to deviate from our normal practice
• Link to article