Advice
Clinical and technical evidence
Clinical and technical evidence
A literature search was carried out for this briefing in accordance with NICE's interim process and methods statement for the production of medtech innovation briefings. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting mibs@nice.org.uk.
Published evidence
There are 6 studies summarised in this briefing, including a total of 1,217 people.
All studies are observational, with 2 being multicentre studies and most studies investigating the prognostic value of the NPi-200. There are further studies that are not summarised here.
The clinical evidence and its strengths and limitations is summarised in the overall assessment of the evidence.
Overall assessment of the evidence
The evidence for the technology is of low methodological quality, and most of the studies are small in terms of patient numbers. For 2 of the studies it was unclear in which country the study took place, and none of the studies were done in the UK. Only 2 studies had a comparator, however they did not report on patient outcomes. The studies show that NPi-200 can predict poor outcomes in critically ill people. Further evidence comparing NPi-200 with standard care, with a large sample size is needed.
Robba et al. (2019)
Study size, design and location
A single-centre prospective observational study of 112 critically ill people.
Intervention and comparator(s)
Neurolight Algiscan (NL) and NPi-200 compared with standard pupillary light reflex (PLR).
Key outcomes
There was a significant correlation between the 2 automated pupillometry devices for pupil size, constriction to light stimulation and constriction velocity, but not for pupillary latency. The NL and the NPi-200 devices' mean bias for pupil size was -0.12 mm (limit of agreement [LoA] -1.29 mm to 1.06 mm), for pupil constriction -1.0% (LoA -9.3% to 7.2%), and for latency 0.02 ms (LoA 0.22 ms to 0.25 ms). There was a significant correlation between pupil size evaluated by clinical examination and by the NL or NPi-200 devices. The mean biases for pupil size measured using NL and NPi-200 and clinical examination were 0.16 mm (LoA -0.99 mm to 1.32 mm) and 0.21 mm (LoA 3.03 mm to 3.30 mm), respectively. Although there was significant correlation between NL and NPi-200 values and clinical examination of the PLR, the 2 devices were not always interchangeable, especially for the evaluation of pupillary latency.
Strengths and limitations
This study compares 2 automated pupillary devices to each other and to standard care (pen torch). The devices were used in random order. This study did not assess the effect of automated pupillary findings on patient outcomes. Furthermore, neurological pupil index (NPi) was not assessed as this was not available in both devices. It is unclear in which country the study took place.
Miroz et al. (2019)
Key outcomes
Non-survivors (n=57) had significantly lower NPi than survivors at all time points (all p<0.01). Abnormal NPi (less than 3, at any time from 24 hours to 72 hours) was 100% specific for 90-day mortality, with no false positives. Adding the 12-hour PREDICT VA-ECMO score to the NPi provided the best prognostic performance (specificity 100%, 95% confidence interval [CI] 92 to 100; sensitivity 60%, 95% CI 46 to 72; area under the receiver operating characteristic curve [AUC] 0.82). Quantitative NPi alone had excellent predictive ability for poor outcome from day 1 after VA-ECMO insertion, with no false positives. Combining NPi and 12-hour PREDICT VA-ECMO score increased the sensitivity of outcome prediction, while maintaining 100% specificity.
Riker et al. (2019)
Key outcomes
All 9 people with 1 or more non-reactive pupil (NPi=0) within 6 hours (range 2 hours) after recovery of spontaneous circulation (ROSC) died, and 86% (12 of 14) with sluggish pupils (NPi less than 3) had poor outcomes. Out of 29 people with normal pupil reactivity (NPi of 3 or more), 15 (52%) had poor outcomes. During targeted temperature management, 95% (20 of 21) of people with non-reactive pupils had poor outcomes, 64% (9 of 14) of people with sluggish pupils had poor outcomes, and 45% (9 of 20 ) of people with normal pupil reactivity had poor outcomes. Pupil size did not predict outcome, but NPi (AUC 0.72 [0.59 to 0.86]; p<0.001), PLR constriction percentage (AUC 0.75 [0.62 to 0.88]; p<0.001) and constriction velocity (AUC 0.78 [0.66 to 0.91]; p<0.001) at 6 hours predicted poor outcome. The best predictor of poor outcome in the first 6 hours after ROSC was an NPi less than 3.7. Very early after resuscitation from cardiac arrest, abnormal NPi and PLR measurements by pupillometer are predictive of poor outcome and are not usually associated with dilated pupils.
Oddo et al. (2018)
Key outcomes
Between day 1 and 3, an NPi of 2 or less had a 51% (95% CI 49 to 53) negative predictive value (NPV) and a 100% positive predictive value (PPV; 0% false-positive rate, 95% CI 0 to 2), with a 100% (95% CI 98 to 100) specificity and 32% (95% CI 27 to 38) sensitivity for the prediction of unfavourable outcome. Using the cut-off of abnormal NPi (less than 3) increased sensitivity (38%, 95% CI 32 to 44) but at the expense of a lower specificity (96%, 95% CI 92 to 98; 6% false-positive rate). Compared with NPi, sPLR had significantly lower PPV and significantly lower specificity (p<0.001 at day 1 and day 2; p=0.06 at day 3). The combination of NPi of 2 or less with bilaterally absent somatosensory evoked potentials (SSEP; n=188 patients) provided higher sensitivity (58% [95% CI 49 to 67] compared with 48% [95% CI 39 to 57] for SSEP alone), with comparable specificity (100% [95% CI 94 to 100]).
Obling et al. (2019)
Key outcomes
Information about 30-day mortality was available for all people in the study. 135 people had OHCA and 51 (38%) people died within 30 days. The median NPi values were 4.10 (interquartile range [IQR] 0.60) in survivors compared to 2.80 (IQR 3.43) in people who did not survive (p<0.0001). Higher NPi values were independently associated with a lower 30-day mortality (odds ratio 0.15, 95% CI 0.06 to 0.29; p<0.0001), and the univariable model had an AUC of 0.87, with a maximal AUC cut-off level for NPi being 3.30 (sensitivity 69% and specificity 93%, PPV 85% and NPV 83%). For people with IHCA and other cardiac diagnoses, they found no association between NPi values and 30-day mortality, and the univariable models showed poor predictive values.
Al-Obaidi et al. (2019)
Key outcomes
Analysis of t-test indicates statistically significant differences for all right and left mean pupilometer values, except right latency (p=0.3000) and repeated measure mixed model (p=0.0001). In people with increased intracranial pressure, mean pupilometer values for left NPi, pupil dilation, pupil size and constriction velocity were lower for both eyes compared with people with normal intracranial pressure. Values were higher in both eyes for people with increased intracranial pressure compared with normal intracranial pressure for right NPi (3.98 and 3.92 respectively; p=0.0300) and left latency (0.27 and 0.25 respectively; p<0.0001). Worsening measures of the PLR using automated pupillometry are associated with elevated intracranial pressure.
Strengths and limitations
The findings from this replication study confirm and extend those of McNett et al. (2018). The registry used in this study was partially funded by NeurOptics, the company that produced the pupillometer.
Recent and ongoing studies
-
Establishing Normative Data for Pupillometer Assessments in Neuro-Intensive Care (ENDPANIC). ClinicalTrials.gov Identifier: NCT02804438. Status: recruiting. Indication: neuro-intensive care unit patients. Devices: NeurOptics NPi-200. Estimated completion date: September 2022. Country: US.
-
Effects of Volatile and Intravenous Anesthetics on Pupillary Function. ClinicalTrials.gov Identifier: NCT03987529. Status: not yet recruiting. Indication: abnormal pupillary functions. Devices: NeurOptics NPi-200. Estimated completion date: February 2021. Country: South Korea.
The company states that it is aware of 2 further UK studies in development.