Article Text


Limitations of sniff nasal pressure in patients with severe neuromuscular weakness
  1. N Hart1,
  2. M I Polkey1,
  3. T Sharshar2,
  4. L Falaize3,
  5. B Fauroux4,
  6. J C Raphaël5,
  7. F Lofaso3
  1. 1Respiratory Muscle Laboratory, Royal Brompton Hospital, Fulham Road, London, UK
  2. 2Service de Réanimation Médicale, Raymond Poincaré Hôpital, Garches, France
  3. 3Service de Physiologie-Explorations Fonctionnelles, Raymond Poincaré Hôpital, Garches, France
  4. 4Paediatric Pulmonary Department, Armand Trousseau Hospital-AP-HP, Paris, France
  5. 5Service de Réanimation Medicale, Raymond Poincaré Hôpital, Garches, France
  1. Correspondence to:
 Professor F Lofaso
 Service de Physiologie-Explorations Fonctionnelles, Raymond Poincaré Hôpital, Garches 92380, France;


Background: Inspiratory muscle strength in patients with neuromuscular disorders can be assessed using sniff inspiratory nasal pressure (Pnsn) and maximum inspiratory mouth pressure (PImax). However, the relative merits of Pnsn against PImax are not known in patients with severe neuromuscular disease.

Objective: To investigate whether severity of disease modifies the relation between Pnsn and PImax.

Methods: Vital capacity (VC), Pnsn, and PImax were measured in 258 patients with neuromuscular disorders.

Results: Data were analysed from 241 patients, 17 being unable to perform PImax or Pnsn manoeuvres. The correlation between Pnsn and PImax was +0.94 (p<0.0001), with a mean (SD) difference between Pnsn and PImax of −4.8 (21.2) cm H2O (the limits of agreement were 37.6 and –47.2 cm H2O). VC (% predicted) was positively correlated with Pnsn/PImax (r  =  +0.86; p<0.0001), with a lower Pnsn/PImax value in patients with a VC <40% of predicted than in those with a VC >40% (0.80 (0.35) v 1.04 (0.41); p<0.0001).

Conclusions: PImax is greater than Pnsn in patients with a severe restrictive ventilatory defect caused by neuromuscular disease. Pnsn may not accurately reflect inspiratory muscle strength in such patients and it is thus advisable to use both tests.

  • neuromuscular disease
  • inspiratory muscle strength
  • FRC, functional residual capacity
  • NIV, non-invasive ventilation
  • PImax, inspiratory mouth pressure
  • Pnsn, sniff inspiratory nasal pressure
  • VC, vital capacity

Statistics from

Although sniff inspiratory nasal pressure (Pnsn) and maximum inspiratory mouth pressure (PImax) are used to assess inspiratory muscle strength in patients with neuromuscular disorders,1,2 it is unclear whether the values for Pnsn are higher than PImax. Stefanutti et al reported that the airway pressure during a sniff manoeuvre was higher than the pressure generated during a maximal static inspiratory manoeuvre,1 whereas a recent study by Iandelli et al suggested the opposite, especially in patients with severe neuromuscular disorders.2 However, in those studies less than 20% of the patients had a vital capacity (VC) below 40% of predicted, which is a level that can be considered to indicate a severe restrictive ventilatory defect (all patients with symptoms of nocturnal hypoventilation and a VC less than 50% of predicted should be considered for long term non-invasive ventilation3). Thus our aim in the current study was to examine the affect of moderate to severe lung restriction on the relation between Pnsn and PImax in patients with neuromuscular disease.


We studied 258 patients with neuromuscular disorders as part of routine clinical evaluation over a three year period at the Raymond Poincaré hospital. Seventeen patients were excluded: nine were unable to perform PImax manoeuvres and eight were unable to perform Pnsn manoeuvres.

All the tests were conducted in a single session, with patients in a seated position. Pnsn and PImax were both measured from functional residual capacity (FRC) in a standard manner, according to previously described methods (PImax is an isometric manoeuvre, while Pnsn is a quasi-isometric manoeuvre).4,5 Pnsn was measured during 10 maximal sniffs,6 while PImax was measured with a flanged mouthpiece with the manoeuvres repeated at least three times or until two identical readings were obtained.7 All signals were measured using a differential pressure transducer (Validyne, Northridge, CA, USA), amplified by a carrier amplifier (Validyne), and passed through an analogue-digital board to a computer running Acqknowledge software (Biopac System, Goleta, CA, USA). Patients received strong verbal encouragement with visual feedback, as previous studies have suggested.8 Spirometry and lung volumes were measured according to standard guidelines and reported as per cent predicted.9 Severity of disease was defined according to VC (% predicted), with a VC <40% being arbitrarily defined as severe restrictive lung disease. PImax and Pnsn are expressed as positive values although they are negative pressures with respect to atmosphere.


All results are expressed as mean (SD). The correlation between Pnsn and PImax was assessed using linear regression analysis. Agreement between Pnsn and PImax was determined using a Bland and Altman plot.10 Pnsn/PImax was used to evaluate the relations between Pnsn, PImax, and VC using linear regression analysis. The difference between Pnsn/PImax in the patients with a VC <40% of predicted and patients with a VC >40% of predicted was assessed using an unpaired t test.


Two hundred and forty one patients with neuromuscular disorders were eligible for analysis (table 1). Their mean (SD) age was 45.1 (16.4) years and 58% were male. No problems were encountered with air leaks during the PImax manoeuvre using the flange mouthpiece. Mean VC (% predicted) was 51.9 (26.0)%, mean PImax was 45.6 (28.2) cm H2O, and mean Pnsn was 41.5 (26.4) cm H2O (35.2% of patients had a VC <40% of predicted and 37.8% were receiving non-invasive ventilation).

Table 1

Patients with neuromuscular disorders eligible for analysis

As expected, there was a positive correlation between Pnsn and PImax (r  =  +0.94, p<0.0001). However, the agreement between Pnsn and PImax, assessed using a Bland and Altman plot10 was relatively poor. Figure 1 shows a Bland and Altman plot with a mean difference between Pnsn and PImax of −4.8 (21.2) cm H2O and limits of agreement of 37.6 and −47.2 cm H2O. Despite this, we observed a positive correlation between VC (% predicted) and Pnsn/PImax (fig 2; r  =  +0.86; p<0.0001), which indicates that as VC falls the absolute value of Pnsn declines more than PImax—that is, PImax is greater than Pnsn. Furthermore, the mean Pnsn/PImax value was lower in patients with a VC <40% of predicted than in patients with a VC greater >40% of predicted (0.86 (0.35) v 1.04 (0.41); p<0.0001). When expressed in absolute values (cm H2O), Pnsn was less than PImax in 147 of the 241 patients.

Figure 1

Bland-Altman plot: difference between sniff inspiratory nasal pressure (Pnsn) and maximum static inspiratory pressures (PImax) plotted against the mean of these two variables (n = 241). Filled symbols, patients with VC <40% of predicted; empty symbols, patients with VC >40% of predicted. VC, vital capacity.

Figure 2

Relation between per cent predicted (% predicted) vital capacity (VC) and sniff inspiratory nasal pressure, normalised for maximum static inspiratory pressure (Pnsn/PImax) (n = 241; r  =  0.86; p<0.0001). Horizontal dashed line indicates VC (% predicted) of 40% and the vertical line indicates unity between Pnsn and PImax (Pnsn/PImax = 1).


As in previous studies of patients with neuromuscular disorders, we found a positive correlation between Pnsn and PImax, but with relatively poor agreement between these two tests.1,2 However, the present findings differ from previous studies1,6 in that the value of PImax was at least the same as or even greater than Pnsn, particularly in those patients with severe ventilatory restriction caused by neuromuscular disease. In fact, the present data are more consistent with, and extend, the observations of Iandelli et al,2 and highlight the limitation of Pnsn in this particular patient population.

Comparing our study with that of Stefanutti et al,1 there are two major differences. The first is that overall Pnsn is less than PImax; the second is that advanced lung restriction from severe neuromuscular disease is associated with a lower value for Pnsn than for PImax. This disparity between the current study and Stefanutti’s could be the result of technical differences, as both Pnsn and PImax manoeuvres in the earlier study were done without visual feedback. As sniffing is a more natural manoeuvre than PImax,11 it is conceivable that without visual feedback Pnsn would be greater than PImax. In addition, in the Stefanutti study there were only three attempts at the unfamiliar PImax manoeuvre, but 10 sniff manoeuvres were done. In our clinical practice we get the patient to do at least three PImax manoeuvres or continue until two identical readings are obtained,7 and our greater PImax value may therefore be attributable to the learning effect. Furthermore, the current study differs from previous studies of patients with amyotrophic lateral sclerosis where facial muscle weakness and leaks around the mouthpiece were a major problem causing a reduction in PImax.12,13

The results of our study are supported by the findings of earlier studies in patients with chronic stable inspiratory muscle weakness2 and in patients with acute respiratory failure,14 showing that Pnsn underestimates inspiratory muscle strength in patients with severe neuromuscular disease. Although we acknowledge that VC is an indicator of both respiratory muscle function and lung and chest wall compliance, it is more commonly used than Pnsn or PImax in clinical practice to monitor sequential changes in respiratory impairment as neuromuscular disease progresses. Furthermore, in patients with advanced restrictive ventilatory defects secondary to neuromuscular disease there are marked reductions in VC with relatively small changes in maximum pressures, owing to the curvilinear relation between VC and maximum inspiratory pressures15—that is, VC is probably a more sensitive marker of disease progression in advanced disease than in mild disease.16

To explain the greater decline of Pnsn than of PImax in patients with severe neuromuscular disease, we must consider the mechanism of sniffing compared with the maximal static inspiratory manoeuvre. Sniff is generated during a ballistic manoeuvre where the inspiratory muscles shorten to a greater extent and at a higher speed than during a PImax manoeuvre, which is a more sustained isometric task. Thus as a result of both the force–velocity and force–length relations in striated muscle, Pnsn should be less than PImax—that is, pressure generation falls with a reduction in the operating length of the muscle and also with an increase in the velocity of muscle shortening. However, as sniffing is more natural than the maximum inspiratory static manoeuvre, it is believed that this accounts for the higher Pnsn than PImax value in normal subjects.6 Also, the value for Pnsn is measured from the peak pressure, whereas PImax is the average pressure sustained over one second, which includes an early peak pressure before falling off to a lower sustained pressure. Thus whichever of these two effects is greatest will influence the value of Pnsn more than PImax. Nevertheless, patients with severe neuromuscular disorders may not be able to perform a rapid sniff manoeuvre owing to significant muscle atrophy. Furthermore, because of the differences in the type of effort and the pattern of muscle activation in the two manoeuvres, the Pnsn value and PImax value probably reflect different aspects of inspiratory muscle function. Thus a more sustained manoeuvre may achieve a greater pattern of inspiratory muscle activation in severely affected patients.


In patients with moderate to severe neuromuscular disease, overall PImax yielded similar values to Pnsn. Although there was a close correlation between Pnsn and PImax, there was a relatively poor agreement between the two variables, and the value for PImax was higher than Pnsn, particularly in patients with a severe restrictive ventilatory defect. Pnsn may overestimate the level of inspiratory muscle weakness, which challenges the view that this is the most appropriate test in this particular group of patients. We therefore suggest that PImax and Pnsn are not interchangeable but are complementary tests and should be used in combination with VC for a complete sequential assessment of inspiratory muscle strength in patients with neuromuscular disease.


NH is funded by a grant from the Association Française Contre Les Myopathies, the Scadding-Morriston Davies respiratory medicine award, and a European Respiratory Society long term fellowship.


View Abstract


  • Competing interests: none declared

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.