Nondepolarizing neuromuscular blocking agents, reversal & Risk of postoperative pneumonia
Anesthesiology 2016; 125: 647-55
Dr John Whittle
Monitoring of neuromuscular blockade is inconsistent & often subjective
The effects of neuromuscular blockade can persist out of the OR
Around 40% of patients entering recovery will have residual block (TOF <0.9)
Ongoing block is associated with increased airway risk/aspiration risk
Respiratory complications, such as pneumonia may result.
Design & Setting
Retrospective database analysis (NSQIP)
University affiliated tertiary care hospital in USA
No intervention (observational)
Surgical cases between 2005 and 2013 in the NSQIP database
30 days post operative follow-up
Excluded automatically if <18 y.o, trauma/transplant, or if operation the complication of another procedure
No change in practice requested.
Whether or not the patient developed postoperative pneumonia (NSQIP definition)
- At least one of CXR + sign of pneumonia (fever, WCC, altered GCS) + either positive micro and/or symptoms
Analysed as matched cohorts as per:
Propensity to receive an intermediate acting non-depolarising NMBA
- i.e. NMBA vs no NMBA
Propensity to receive reversal if NMBA given
13,920 cases in the database over the time period:
1455 matched who had NMBA to 1455 who did not
Among the 10,594 who did receive an NMBA ( the majority!)
1320 matched who got reversal to 1320 who did not
Patients receiving an NMBA had a higher incidence rate of pneumonia (9 vs 5 person days at risk) (IRR 1.79 (1.08 to 3.07)
Patients who were not reversed were 2.26 times more likely to develop pneumonia
(IRR 2.26 (1.65-3.03)
An association was found between use of NMBAs and postoperative pneumonia
Novel data is presented regarding reversal of NMBA’s and the risk of developing pneumonia
More viglilance is required with use of NMBAs
1. Large Cohort
2. Sound statistical methods – strong attempts made to reduce bias
3. Relevant clinical question associated with a potential mechanism
4. Sound end-point (pneumonia)
5. ‘Real world’ environment
1. Observational and therefore cannot establish causality
2. Cannot rule out possibility of bias from unmeasured confounders
3. Diagnosis of pneumonia
4. Risk of misclassification in database
5. Older/Sicker patients may be more likely to undergo investigation
6. Generalisability may be an issue as single centre tertiary environment
7. No TOF data included
8. Recommend a prospective clinical trial
9. Issues with factors used for propensity matching: no predictive scores for pulmonary complications, not clear how in reality how clinically well matched the patients were in terms of underlying medical conditions and clinical indications for intubation.
This is a major limitation of this study at first glance: normally there is a clinical reason for endotracheal intubation (the most common association with neuromuscular blockade), the primary outcome of this study is therefore massively influenced by this.
Propensity score matched analysis used to overcome this; reasonable choices made, however:
Scores were limited to that which was coded in the database – no specific analysis of aspiration risk or PPC risk scores etc.
Surely propensity scoring will eliminate many patients undergoing longer surgery since it is highly unlikely that they will be intubated without NMB: though 98% of matches reported as being 2 digit.. this is not explicitly reported in this paper
10. There was no pre-study power analysis. The post-hoc analysis reported appears to support numbers used
11. The major issue with this study lies not around the statistical tests used – there was major effort made to reduce bias, but rather the data used to form the statistical models. Inadequate or irrelevant data into a model will result in unrepresentative results being produced. Whilst most of the data put into this model was completely relevant, several key missing features may skew the results.
More objective evidence is needed to determine causation, however this paper would support a move towards further work focusing on quantitative train of four monitoring and implications of inadequate blockade on the likelihood of developing postoperative pneumonia. Why patients who did not receive neuromuscular blocking agents are less likely to develop pneumonia is not clear from this paper. It is possible that processes around intubation, as well as the clinical reasons for intubation and patient morbidity factors directly contributing to increased risk of postoperative pneumonia are as if not more important.
Potential for impact
Whilst no conclusive mechanistic evidence is presented in this paper for causation of pneumonia due to residual neuromuscular blockade, the data strongly supports a re evaluation of how meticulous we are in ensuring adequate reversal at the end of surgery.