Use of Sniff Nasal-Inspiratory Force to Predict Survival in Amyotrophic Lateral Sclerosis
http://www.100md.com
美国呼吸和危急护理医学 2005年第2期
Departments of Medicine, Neurology
Epidemiology, Royal College of Surgeons in Ireland, Beaumont Hospital, Dublin, Ireland
ABSTRACT
Respiratory muscle weakness is the usual cause of death in amyotrophic lateral sclerosis. The prognostic value of the forced vital capacity (FVC), mouth-inspiratory force, and sniff nasal-inspiratory force were established in a group of 98 patients with amyotrophic lateral sclerosis who were followed trimonthly for 3 years. Sniff nasal-inspiratory force correlated with the transdiaphragmatic pressure (r = 0.9, p < 0.01). Sniff nasal-inspiratory force was most likely to be recorded at the last visit (96% of cases), compared with either the FVC or mouth-inspiratory force (86% and 81%, respectively, p < 0.01). A sniff nasal-inspiratory force less than 40 cm H2O was significantly related with nocturnal hypoxemia. When sniff nasal-inspiratory force was less than 40 cm H2O, the hazard ratio for death was 9.1 (p = 0.001), and the median survival was 6 ± 0.3 months. The sensitivity of FVC < 50% for predicting 6-month mortality was 58% with a specificity of 96%, whereas sniff nasal-inspiratory force less than 40 H2O had a sensitivity of 97% and a specificity of 79% for death within 6 months. Thus the sniff nasal-inspiratory force test is a good measure of respiratory muscle strength in amyotrophic lateral sclerosis, it can be performed by patients with advanced disease, and it gives prognostic information.
Key Words: noninvasive ventilation respiratory failure respiratory muscle strength
Amyotrophic lateral sclerosis (ALS) is a common neurologic disease associated with an inexorable decline in muscle strength (1). Most patients with ALS die of respiratory failure resulting from respiratory muscle weakness. A forced vital capacity (FVC) of less than 50% of the predicted value has been shown to be associated with a poor prognosis (2, 3). An FVC less than 50% has since been used as an important endpoint in many clinical trials and decision-making events for patients with ALS.
However, there are several reasons why FVC is not an ideal test of respiratory muscle strength in ALS. The FVC may not fall until there is profound muscle weakness, a consequence of the sigmoid relationship of the lung pressureeCvolume curves (4). Second, patients with bulbar muscle weakness cannot make a tight seal around the mouthpiece; therefore, recorded values may not reflect the patient's true respiratory muscle strength. Tests of respiratory muscle strength such as the maximal mouth-inspiratory force (MIF) and maximal mouth-expiratory force are related to survival in ALS. However, these tests are also difficult to perform when the bulbar muscles are affected. The lack of a single test with a predictive value for mortality in ALS has been emphasized in a prior evidence-based review (5).
A sniff is a short, sharp, voluntary inspiratory maneuver. Prior studies have shown that the maximal sniff nasal-inspiratory force (SNIF) correlates well with invasive and nonvolitional tests of diaphragm strength (6). Prior studies have identified that the SNIF is sensitive to changes in respiratory muscle strength (7). Thus the SNIF test may be a good way to monitor respiratory muscle strength as the disease progresses. Furthermore, using these data, it should be possible to establish a predictive relationship between SNIF and survival in patients with ALS. Hence, the hypothesis of this study was that SNIF values could be related to the prognosis of ALS. To do this, we prospectively recorded measures of respiratory muscle function over a 3-year period in consecutive patients referred to a specialist ALS clinic. The results of these studies have already been presented in abstract form (8).
METHODS
Ninety-eight patients diagnosed with ALS by El Escorial criteria (9) were studied trimonthly for 3 years at a specialist ALS clinic. The clinic is staffed by a neurologist, specialist nurse, speech and language therapist, occupational therapist, dietician, and respiratory technician. No patients declined to be enrolled and none were enrolled in other clinical trials of potentially disease-modifying drugs during the course of this study. All patients were prescribed riluzole. No subject had received ventilatory support and none had experienced an adverse respiratory event before entry into the study. Because poor attendance by patients with incapacity from advanced disease may give a bias toward more mobile subjects, regular review at this clinic was encouraged by a liaison nurse and follow-up phone conversations with family members. Noninvasive positive pressure ventilation was recommended for patients with respiratory failure (PaCO2 > 6.5 Kpa in the morning or mean oxygen levels < 90% overnight) or an FVC < 50%. The cause and date of death were obtained by examination of the death certificate, and independent verification that the death was from respiratory failure was obtained from a community-based liaison nurse who visited the patients during the terminal phase of the illness. Subjects were divided into bulbar- and nonbulbar-onset ALS. Bulbar-onset ALS referred to the population of patients who presented with upper and lower motor neuron signs in the bulbar region as their first reported symptoms. Verbal informed consent was obtained from all subjects. The hospital's ethics committee approved the study.
Spirometry
The same experienced qualified respiratory technician performed the tests, and subjects were given adequate time to rest between the tests. At each visit, the FVC was recorded with a Microlab spirometer (MicroMedical Limited, Rochester, Kent, UK) using standards recommended by the American Thoracic Society. The best of three reproducible values was recorded. Where necessary, the maximal arm span was used to establish a percent predicted value.
Recording MIF and Maximal Expiratory Force
MIF and maximal expiratory force (MEF) were measured using a hand-held pressure meter with a flanged mouthpiece (MicroMedical Limited). The subject was instructed to suck in/blow out as hard as possible and the best of six tests of at least 1-second duration were recorded.
Recording Nasal Inspiratory and Expiratory Force
At the start of this study, a commercial nasal inspiratory pressure meter was not available, so the device used for MIF measurements was modified for SNIF measurements. A plug was inserted into a nostril and the center of this plug snugly held the tip of a polyethylene catheter (Intersurgical Scientific Instruments, Oxford, UK) within the nasal cavity. This polyethylene catheter was 30 cm in length and had an internal diameter of 2 mm and the other end was attached to the hand-held pressure transducer (a photomicrograph of the device is shown in Figure E1 in the online supplement). The contralateral nose was occluded and the patients were instructed to close the mouth and to breathe out and then to take a deep sniff or a maximal inspiratory effort from end-expiratory lung volume or to breathe in and then to make an expiratory effort from end-inspiratory lung volume. Because a subject's nasal passages may be blocked, recordings were taken from both nostrils (left side first) and the highest of six recorded values sustained for over 1 second was recorded. An initial validation study was performed in 102 healthy control subjects (age range 20eC86; mean age 46.3 years).
Recording Oxygen Levels During Sleep
A domiciliary watch oximetry (Minolta Pulsox-3 Oxygen situation monitor, Stowood Scientific Instruments, Oxford, UK) was used to record oxygen saturations. The device recorded oxygen saturation levels at 4-second intervals. The device's computer software (Minolta Pulsox, version 3.02, Stowood Scientific Instruments) was used to remove obvious errors from device displacement and to obtain the mean oxygen saturations and the proportion of recording time spent with oxygen saturation less than 90%.
Transdiaphragmatic Pressure
Esophageal pressures (Pes) and gastric pressure (Pgas) were recorded using a Grass polygraph (Grass Telefactor; Astro-Med, Inc., West Warwick, RI); transdiaphragmatic pressure (Pdi) was calculated in a subgroup of 24 subjects. After calibration, a silastic tube with two separate air-filled pressure probes was inserted into one nostril. The subjects were instructed to swallow and the tube was passed through the oropharynx with each swallow. Direct visualization of the pressure tracing produced with each sniff was used to position the tube so that one probe recorded gastric pressure (positive deflection during inspiration) and the other recorded esophageal pressure (negative deflection during inspiration). Subjects were then instructed to sniff, and measures of esophageal pressure and transdiaphragmatic pressure were recorded (gastric pressure, esophageal pressure).
Statistical Analysis
Statistical analysis was performed using StataSE release 8.1. Data were expressed as event history data, which allows participants to be switched between at-risk groups on the basis of their most recently measured clinical parameters. Where clinical measurements could not be made at a particular follow-up, values were carried forward until the next measurement. Analysis was by most recent value of clinical parameters. Cox regression was used to model the prognostic effect of the clinical parameters studied. To examine the relationship between different noninvasive measurements of respiratory muscle strength and transdiaphragmatic pressure were calculated using Lin's concordance coefficient (10, 11). This is a measure of agreement that is scaled between zero and one. To overcome differences of scale, all measures were transformed by ranking them. The concordance coefficient therefore represents the agreement in rank order between each of the measures and transdiaphragmatic pressure. Kaplan-Meier survival was calculated and used to identify survival quantiles.
RESULTS
The demographic details of the subjects, at the time of entry into the study are given in Table 1. Ninety-eight patients were studied, and over the 3 years, 39 of these died. The mean duration of symptoms before enrollment was 11.7 months, and 13 patients required a wheelchair for mobility at the point of first recording of respiratory muscle strength. At the time of entry into the study, the mean FVC was 78 ± 25%, and the SNIF and MIF were also diminished at 51.6 ± 21.7 and 47.3 ± 22.4 cm H2O. Over the study period, there were 271 visits, and 71 patients had more than two recordings performed. Measures of respiratory muscle strength were obtained in 87% of subjects within 3 months of death and were obtained in 97% within 6 months of death. During the course of the study, 20 subjects were started on noninvasive ventilation, but no subject was placed on mechanical ventilation. Of these, eight tolerated the device and used it at home. The mean survival for this group was 4.2 ± 4.3 months, and 2.2 ± 1.4 months for the group who did not use the device.
SNIF Correlates Well with Invasive Measurement of Respiratory Muscle Strength
An initial validation study was performed in 102 healthy control subjects (age range 20eC86, mean age 46.3 years). The mean SNIF was 90 ± 31 cm H2O; for SNEF, it was 134 ± 48 cm H2O. Transdiaphragmatic pressure was assessed in a subgroup of 24 subjects, 12 of whom had bulbar disease. A good concordance of SNIF to transdiaphragmatic pressure was seen (rho_c = 0.661) (Figure 1 and Table 2).
SNIF is a Reproducible Test in Advanced ALS
The ability of each of the 98 subjects to perform a reproducible test of each of the measures of respiratory strength at their last visit before death or study completion is shown in Figure 2. The SNIF was significantly more likely to be recorded than either the FVC or MIF (p < 0.01; Figure 2). Among all subjects, only four (4%) were unable to record a SNIF; two of these subjects had severe bulbar weakness: one had a deviated nasal septum and one subject's anterior nares was too large to retain the nasal plug during the maneuver. Fourteen subjects (14%), all with significant bulbar dysfunction, were unable to perform a reproducible FVC recording. MIF could not be obtained in 19 (19%) subjects, all of whom had bulbar weakness. These subjects all reported difficulty in performing the test against a closed airway, a commonly identified problem with the MIF test, even in individuals with normal muscle function (4, 12).
Relationship Between Overnight Oximetry and FVC, MIF, and SNIF
Fifty-seven subjects had overnight oximetry performed, although not all of these subjects could perform all of the measures of respiratory muscle strength. Correlation analysis of the mean nocturnal oxygen saturation levels to SNIF was r = 0.4, p = 0.001; for MIF, r = 0.322, p = 0.056; and for FVC, r = 0.39, p = 0.003 (Figure 3).
Threshold Levels of FVC, MIF, and SNIF to Predict Nocturnal Hypoxemia and Death
To evaluate the clinical relevance of the noninvasive tests for predicting mortality, we used threshold levels for each test. The association between FVC, MIF, and SNIF categories with time spent at less than 90% oxygen saturation at night are shown in Figure 4. Only SNIF category (above or below 40 cm H2O) was significantly associated with desaturation on this analysis (p = 0.01).
Predictors of Mortality in ALS
Each test was considered as a continuous variable, and Cox regression analysis was used to examine the hazard ratios for predicting death. Table E1 (in the online supplement) shows the results of this analysis. Age, body mass index, FVC (% predicted), SNIF, MIF, and SNEF were all independently associated with a significantly increased risk of death at any point of follow-up. In addition, the oximetry variables measured; mean O2 and proportion of night spent at less than 90% O2 saturation were related to death.
A SNIF less than 40 cm H2O was associated with a hazard risk for death of 9.1 (95% CI 4eC20.8, p < 0.001) (Figure E2). The 25th percentile for mortality when the SNIF fell to less than 40 cm H2O was 3.46 ± 0.1 months (95% CI 2.51eC5.52) and the median (50%) mortality was 6 ± 0.3 months (95% CI 2.51eC8.45). When the eight patients who used noninvasive ventilation were excluded from the data set, there was still no effect on the major result of the study (i.e., that SNIF < 40 cm H2O was a positive predictor of death; data not shown). Figure 5 shows the Kaplan-Meier survival curve for all subjects separated into change of SNIF of 10 cm H2O. The hazard ratio for death when the SNIF was less than 30 cm H2O compared with those with a SNIF more than 30 cm H2O was 5.9 (95% CI 3eC12). When FVC fell to less than 50%, the hazard ratio for death was 5.66 (95% CI 2.73eC11.73, p < 0.001).
Among patients with a SNIF less than 40 cm H2O, 66% had an FVC greater than 50%; in this group, the hazard ratio for death was 13.6 (95% CI 3.1eC54.7, p < 0.001) (Table 3). We defined sensitivity as the proportion of subjects who died when their measurements were FVC less than 50% and MIF and SNIF less than 40 cm H2O 6 months before death and the specificity as the proportion of subjects who lived for more than 6 months when their test was above the cutoff value. The sensitivity of FVC less than 50% for predicting 6-month mortality was 58% with a specificity of 96%, whereas for SNIF less than 40 cm H2O, the test had a sensitivity of 97% and a specificity of 79% for death within 6 months (Table 4).
DISCUSSION
The results of this study show that SNIF correlated well with transdiaphragmatic strength and so is a good measure of respiratory muscle strength. Furthermore, at the later stages of the disease, the SNIF test could still be performed by 96% of the patients. Finally, we established that a SNIF less than 40 cm H2O is associated with a median survival of 6 months. Thus the SNIF test is a sensitive means of detecting respiratory muscle strength, it can be recorded in advanced disease, and it gives important prognostic information.
Prior studies of the natural history of ALS have identified that deteriorations in pulmonary function tests are the most important prognostic factors in survival (2, 3, 13, 14). These studies have generally used the FVC as the main measure of respiratory muscle strength. The FVC is not a particularly good measure of respiratory muscle strength because the shape of the pressureeCvolume curve means that it does not detect modest falls in muscle strength. Thus alternative measures of muscle strength with MIF and SNIF testing have been investigated in some prior studies. One retrospective study demonstrated an association between MIF and survival, but did not establish a value that gave a guide to prognosis (15). Prior prospective studies have reported that composite respiratory scores—a combination of FVC and MIF—were related to survival. However, in these studies too, a threshold value for either test which provided prognostic information was not established (3, 14).
The study population was prospectively recruited and followed over 3 years. Approximately 40% of the study subjects were women, reflecting the expected male:female ratio of 1.6:1 in an ALS population. Although the average FVC was slightly reduced at the start of the study, the direct measures of muscle strength, MIF and SNIF, were considerably reduced in most patients. At the start of this study, a device to measure SNIF was not commercially available, so we modified a hand-held mouth pressure meter to record these pressures. To establish the reliability of this new device, we undertook a series of validation studies, testing the fidelity of the recording device and establishing a range of values of SNIF in a cohort of healthy volunteers of a similar age as the ALS patients. The mean value of SNIF was 90 cm H2O, which is in agreement with prior data, suggesting that our modified device was suitable for this study (16). In addition, we related SNIF, MIF, and FVC to transdiaphragmatic muscle strength in a subgroup of patients with ALS. In these studies, it was shown that both MIF and SNIF correlated better with transdiaphragmatic pressure than with FVC. This confirms the results of prior studies indicating that MIF and SNIF are better methods of recording respiratory muscle strength than FVC. We had noticed, clinically, that weakness of the bulbar muscle often prevents patients from making a complete seal around a mouth device; consequently, FVC and MIF may not be recorded accurately at a time when the information is most relevant. In the current study, 4 patients (4%) could not perform SNIF at the last visit, whereas 14 (14%) and 19 (19%) could not perform the FVC and MIF, respectively. Thus, although MIF and SNIF are better tests of respiratory muscle strength than FVC, more patients with advanced disease can perform the SNIF test, making this the preferred test to measure respiratory muscle strength in ALS.
The primary purpose of this study was to establish if a SNIF value could be established that predicted mortality in patients with ALS. We categorized SNIF measurements into units of 10 cm H2O below 50 cm H2O and related these to mortality. A SNIF less than 50 cm H2O had a relatively good prognosis, whereas a SNIF less than 40 cm H2O had a median survival of less than 6 months and a 1-year survival rate less than 25%. When the SNIF fell below 30 cm H2O, the median survival was 3 months; few patients with such a measure survived more than 6 months.
Although the primary end point of the study was mortality, we also related the tests of lung function to nocturnal oxygen saturation. Nocturnal hypoxia is a feature of hypoventilation in people with otherwise healthy lungs and is seen before daytime hypoxia and so reflects early respiratory muscle weakness. A SNIF less than 40 cm H2O was the only test that was significantly correlated with nocturnal hypoxia. Thus a fall in SNIF to this level may be an appropriate time to consider noninvasive ventilation. Studies in our unit are in progress to establish if initiation of nocturnal ventilatory support leads to increased survival when the SNIF falls to this level.
Noninvasive ventilation may prolong life and improve quality of life in patients with ALS (17eC19). However, there is no consensus on when to initiate this treatment (20). The timing of instituting ventilatory support and discussion of the overall prognosis are important issues for patients. The prognosis, the recommended timing of noninvasive ventilation, and assessment of new treatments in ALS all use FVC% as the standard measure of respiratory performance (21, 22). The results of this study highlight the limitations of using the FVC in ALS. First, it cannot be obtained in about 20% of subjects at the later stages of the disease, and second, it is not sensitive to important changes in respiratory muscle strength because 75% of subjects with important levels of muscle weakness (SNIF < 40 cmH2O) still had an FVC greater than 50%. This is important because, in this group, the hazard ratio for death within 6 months was 13.6. In summary, the SNIF is a valuable tool to monitor respiratory muscle strength in patients with ALS.
Acknowledgments
The authors are grateful to the patients who participated in this study.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
These authors contributed equally to this manuscript.
REFERENCES
Kreitzer S, Saunders NA, Tyler R, Ingram RH Jr. Respiratory muscle function in amyotrophic lateral sclerosis. Am Rev Respir Dis 1978;117:437eC447.
Fallat RJ, Jewitt B, Bass M, Kamm B, Norris FH Jr. Spirometry in amyotrophic lateral sclerosis. Arch Neurol 1979;36:74eC80.
Stambler N, Charatan M, Cedarbaum JM. Prognostic indicators of survival in ALS. ALS CNTF Treatment Study Group. Neurology 1998;50:66eC72.
Fitting JW, Paillex R, Hirt L, Aebischer P, Schluep M. Sniff nasal pressure: a sensitive respiratory test to assess progression of amyotrophic lateral sclerosis. Ann Neurol 1999;46:887eC893.
Miller RG, Rosenberg JA, Gelinas DF, Mitsumoto H, Newman D, Sufit R, Borasio GD, Bradley WG, Bromberg MB, Brooks BR, et al. Practice parameter: the care of the patient with amyotrophic lateral sclerosis (an evidence-based review). Muscle Nerve 1999;22:1104eC1118.
Lyall RA, Donaldson N, Polkey MI, Leigh PN, Moxham J. Respiratory muscle strength and ventilatory failure in amyotrophic lateral sclerosis. Brain 2001;124:2000eC2013.
Stefanutti D, Benoist MR, Scheinmann P, Chaussain M, Fitting JW. Usefulness of sniff nasal pressure in patients with neuromuscular or skeletal disorders. Am J Respir Crit Care Med 2000;162:1507eC1511.
Morgan RK, McNally S, Hardiman O, Costello RW. SNIPmax as a predictor of mortality in amyotrophic lateral sclerosis. Am J Respir Crit Care Med 2004;169:A440.
Bradley WG. Overview of motor neuron disease: classification and nomenclature. Clin Neurosci 1995;3:323eC326.
Lin LI-K. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989;45:255eC268.
Lin LI-K. A note on the concordance correlation coefficient. Biometrics 2000;56:324eC325.
Polkey MI, Green M, Moxham J. Measurement of respiratory muscle strength. Thorax 1995;50:1131eC1135.
Haverkamp LJ, Appel V, Appel SH. Natural history of amyotrophic lateral sclerosis in a database population: validation of a scoring system and a model for survival prediction. Brain 1995;118:707eC719.
Ringel SP, Murphy JR, Alderson MK, Bryan W, England JD, Miller RG, Petajan JH, Smith SA, Roelofs RI, Ziter F, et al. The natural history of amyotrophic lateral sclerosis. Neurology 1993;43:1316eC1322.
Gay PC, Westbrook PR, Daube JR, Litchy WJ, Windebank AJ, Iverson R. Effects of alterations in pulmonary function and sleep variables on survival in patients with amyotrophic lateral sclerosis. Mayo Clin Proc 1991;66:686eC694.
Uldry C, Fitting JW. Maximal values of sniff nasal inspiratory pressure in healthy subjects. Thorax 1995;50:371eC375.
Newsom-Davis IC, Lyall RA, Leigh PN, Moxham J, Goldstein LH. The effect of non-invasive positive pressure ventilation (NIPPV) on cognitive function in amyotrophic lateral sclerosis (ALS): a prospective study. J Neurol Neurosurg Psychiatry 2001;71:482eC487.
Aboussouan LS, Khan SU, Meeker DP, Stelmach K, Mitsumoto H. Effect of noninvasive positive-pressure ventilation on survival in amyotrophic lateral sclerosis. Ann Intern Med 1997;127:450eC453.
Pinto AC, Evangelista T, Carvalho M, Alves MA, Sales Luis ML. Respiratory assistance with a non-invasive ventilator (BiPAP) in MND/ALS patients: survival rates in a controlled trial. J Neurol Sci 1995;129:19eC26.
Lechtzin N, Rothstein J, Clawson L, Diette GB, Wiener CM. Amyotrophic lateral sclerosis: evaluation and treatment of respiratory impairment. Amyotroph Lateral Scler Other Motor Neuron Disord 2002;3:5eC13.
Cudkowicz ME, Shefner JM, Schoenfeld DA, Brown Jr RH Jr, Johnson H, Qureshi M, Jacobs M, Rothstein JD, Appel SH, Pascuzzi RM, et al. A randomized, placebo-controlled trial of topiramate in amyotrophic lateral sclerosis. Neurology 2003;61:456eC464.
Groeneveld GJ, Veldink JH, van der Tweel I, Kalmijn S, Beijer C, de Visser M, Wokke JH, Franssen H, van den Berg LH. A randomized sequential trial of creatine in amyotrophic lateral sclerosis. Ann Neurol 2003;53:437eC445.(Ross K. Morgan, Stephen M)
Epidemiology, Royal College of Surgeons in Ireland, Beaumont Hospital, Dublin, Ireland
ABSTRACT
Respiratory muscle weakness is the usual cause of death in amyotrophic lateral sclerosis. The prognostic value of the forced vital capacity (FVC), mouth-inspiratory force, and sniff nasal-inspiratory force were established in a group of 98 patients with amyotrophic lateral sclerosis who were followed trimonthly for 3 years. Sniff nasal-inspiratory force correlated with the transdiaphragmatic pressure (r = 0.9, p < 0.01). Sniff nasal-inspiratory force was most likely to be recorded at the last visit (96% of cases), compared with either the FVC or mouth-inspiratory force (86% and 81%, respectively, p < 0.01). A sniff nasal-inspiratory force less than 40 cm H2O was significantly related with nocturnal hypoxemia. When sniff nasal-inspiratory force was less than 40 cm H2O, the hazard ratio for death was 9.1 (p = 0.001), and the median survival was 6 ± 0.3 months. The sensitivity of FVC < 50% for predicting 6-month mortality was 58% with a specificity of 96%, whereas sniff nasal-inspiratory force less than 40 H2O had a sensitivity of 97% and a specificity of 79% for death within 6 months. Thus the sniff nasal-inspiratory force test is a good measure of respiratory muscle strength in amyotrophic lateral sclerosis, it can be performed by patients with advanced disease, and it gives prognostic information.
Key Words: noninvasive ventilation respiratory failure respiratory muscle strength
Amyotrophic lateral sclerosis (ALS) is a common neurologic disease associated with an inexorable decline in muscle strength (1). Most patients with ALS die of respiratory failure resulting from respiratory muscle weakness. A forced vital capacity (FVC) of less than 50% of the predicted value has been shown to be associated with a poor prognosis (2, 3). An FVC less than 50% has since been used as an important endpoint in many clinical trials and decision-making events for patients with ALS.
However, there are several reasons why FVC is not an ideal test of respiratory muscle strength in ALS. The FVC may not fall until there is profound muscle weakness, a consequence of the sigmoid relationship of the lung pressureeCvolume curves (4). Second, patients with bulbar muscle weakness cannot make a tight seal around the mouthpiece; therefore, recorded values may not reflect the patient's true respiratory muscle strength. Tests of respiratory muscle strength such as the maximal mouth-inspiratory force (MIF) and maximal mouth-expiratory force are related to survival in ALS. However, these tests are also difficult to perform when the bulbar muscles are affected. The lack of a single test with a predictive value for mortality in ALS has been emphasized in a prior evidence-based review (5).
A sniff is a short, sharp, voluntary inspiratory maneuver. Prior studies have shown that the maximal sniff nasal-inspiratory force (SNIF) correlates well with invasive and nonvolitional tests of diaphragm strength (6). Prior studies have identified that the SNIF is sensitive to changes in respiratory muscle strength (7). Thus the SNIF test may be a good way to monitor respiratory muscle strength as the disease progresses. Furthermore, using these data, it should be possible to establish a predictive relationship between SNIF and survival in patients with ALS. Hence, the hypothesis of this study was that SNIF values could be related to the prognosis of ALS. To do this, we prospectively recorded measures of respiratory muscle function over a 3-year period in consecutive patients referred to a specialist ALS clinic. The results of these studies have already been presented in abstract form (8).
METHODS
Ninety-eight patients diagnosed with ALS by El Escorial criteria (9) were studied trimonthly for 3 years at a specialist ALS clinic. The clinic is staffed by a neurologist, specialist nurse, speech and language therapist, occupational therapist, dietician, and respiratory technician. No patients declined to be enrolled and none were enrolled in other clinical trials of potentially disease-modifying drugs during the course of this study. All patients were prescribed riluzole. No subject had received ventilatory support and none had experienced an adverse respiratory event before entry into the study. Because poor attendance by patients with incapacity from advanced disease may give a bias toward more mobile subjects, regular review at this clinic was encouraged by a liaison nurse and follow-up phone conversations with family members. Noninvasive positive pressure ventilation was recommended for patients with respiratory failure (PaCO2 > 6.5 Kpa in the morning or mean oxygen levels < 90% overnight) or an FVC < 50%. The cause and date of death were obtained by examination of the death certificate, and independent verification that the death was from respiratory failure was obtained from a community-based liaison nurse who visited the patients during the terminal phase of the illness. Subjects were divided into bulbar- and nonbulbar-onset ALS. Bulbar-onset ALS referred to the population of patients who presented with upper and lower motor neuron signs in the bulbar region as their first reported symptoms. Verbal informed consent was obtained from all subjects. The hospital's ethics committee approved the study.
Spirometry
The same experienced qualified respiratory technician performed the tests, and subjects were given adequate time to rest between the tests. At each visit, the FVC was recorded with a Microlab spirometer (MicroMedical Limited, Rochester, Kent, UK) using standards recommended by the American Thoracic Society. The best of three reproducible values was recorded. Where necessary, the maximal arm span was used to establish a percent predicted value.
Recording MIF and Maximal Expiratory Force
MIF and maximal expiratory force (MEF) were measured using a hand-held pressure meter with a flanged mouthpiece (MicroMedical Limited). The subject was instructed to suck in/blow out as hard as possible and the best of six tests of at least 1-second duration were recorded.
Recording Nasal Inspiratory and Expiratory Force
At the start of this study, a commercial nasal inspiratory pressure meter was not available, so the device used for MIF measurements was modified for SNIF measurements. A plug was inserted into a nostril and the center of this plug snugly held the tip of a polyethylene catheter (Intersurgical Scientific Instruments, Oxford, UK) within the nasal cavity. This polyethylene catheter was 30 cm in length and had an internal diameter of 2 mm and the other end was attached to the hand-held pressure transducer (a photomicrograph of the device is shown in Figure E1 in the online supplement). The contralateral nose was occluded and the patients were instructed to close the mouth and to breathe out and then to take a deep sniff or a maximal inspiratory effort from end-expiratory lung volume or to breathe in and then to make an expiratory effort from end-inspiratory lung volume. Because a subject's nasal passages may be blocked, recordings were taken from both nostrils (left side first) and the highest of six recorded values sustained for over 1 second was recorded. An initial validation study was performed in 102 healthy control subjects (age range 20eC86; mean age 46.3 years).
Recording Oxygen Levels During Sleep
A domiciliary watch oximetry (Minolta Pulsox-3 Oxygen situation monitor, Stowood Scientific Instruments, Oxford, UK) was used to record oxygen saturations. The device recorded oxygen saturation levels at 4-second intervals. The device's computer software (Minolta Pulsox, version 3.02, Stowood Scientific Instruments) was used to remove obvious errors from device displacement and to obtain the mean oxygen saturations and the proportion of recording time spent with oxygen saturation less than 90%.
Transdiaphragmatic Pressure
Esophageal pressures (Pes) and gastric pressure (Pgas) were recorded using a Grass polygraph (Grass Telefactor; Astro-Med, Inc., West Warwick, RI); transdiaphragmatic pressure (Pdi) was calculated in a subgroup of 24 subjects. After calibration, a silastic tube with two separate air-filled pressure probes was inserted into one nostril. The subjects were instructed to swallow and the tube was passed through the oropharynx with each swallow. Direct visualization of the pressure tracing produced with each sniff was used to position the tube so that one probe recorded gastric pressure (positive deflection during inspiration) and the other recorded esophageal pressure (negative deflection during inspiration). Subjects were then instructed to sniff, and measures of esophageal pressure and transdiaphragmatic pressure were recorded (gastric pressure, esophageal pressure).
Statistical Analysis
Statistical analysis was performed using StataSE release 8.1. Data were expressed as event history data, which allows participants to be switched between at-risk groups on the basis of their most recently measured clinical parameters. Where clinical measurements could not be made at a particular follow-up, values were carried forward until the next measurement. Analysis was by most recent value of clinical parameters. Cox regression was used to model the prognostic effect of the clinical parameters studied. To examine the relationship between different noninvasive measurements of respiratory muscle strength and transdiaphragmatic pressure were calculated using Lin's concordance coefficient (10, 11). This is a measure of agreement that is scaled between zero and one. To overcome differences of scale, all measures were transformed by ranking them. The concordance coefficient therefore represents the agreement in rank order between each of the measures and transdiaphragmatic pressure. Kaplan-Meier survival was calculated and used to identify survival quantiles.
RESULTS
The demographic details of the subjects, at the time of entry into the study are given in Table 1. Ninety-eight patients were studied, and over the 3 years, 39 of these died. The mean duration of symptoms before enrollment was 11.7 months, and 13 patients required a wheelchair for mobility at the point of first recording of respiratory muscle strength. At the time of entry into the study, the mean FVC was 78 ± 25%, and the SNIF and MIF were also diminished at 51.6 ± 21.7 and 47.3 ± 22.4 cm H2O. Over the study period, there were 271 visits, and 71 patients had more than two recordings performed. Measures of respiratory muscle strength were obtained in 87% of subjects within 3 months of death and were obtained in 97% within 6 months of death. During the course of the study, 20 subjects were started on noninvasive ventilation, but no subject was placed on mechanical ventilation. Of these, eight tolerated the device and used it at home. The mean survival for this group was 4.2 ± 4.3 months, and 2.2 ± 1.4 months for the group who did not use the device.
SNIF Correlates Well with Invasive Measurement of Respiratory Muscle Strength
An initial validation study was performed in 102 healthy control subjects (age range 20eC86, mean age 46.3 years). The mean SNIF was 90 ± 31 cm H2O; for SNEF, it was 134 ± 48 cm H2O. Transdiaphragmatic pressure was assessed in a subgroup of 24 subjects, 12 of whom had bulbar disease. A good concordance of SNIF to transdiaphragmatic pressure was seen (rho_c = 0.661) (Figure 1 and Table 2).
SNIF is a Reproducible Test in Advanced ALS
The ability of each of the 98 subjects to perform a reproducible test of each of the measures of respiratory strength at their last visit before death or study completion is shown in Figure 2. The SNIF was significantly more likely to be recorded than either the FVC or MIF (p < 0.01; Figure 2). Among all subjects, only four (4%) were unable to record a SNIF; two of these subjects had severe bulbar weakness: one had a deviated nasal septum and one subject's anterior nares was too large to retain the nasal plug during the maneuver. Fourteen subjects (14%), all with significant bulbar dysfunction, were unable to perform a reproducible FVC recording. MIF could not be obtained in 19 (19%) subjects, all of whom had bulbar weakness. These subjects all reported difficulty in performing the test against a closed airway, a commonly identified problem with the MIF test, even in individuals with normal muscle function (4, 12).
Relationship Between Overnight Oximetry and FVC, MIF, and SNIF
Fifty-seven subjects had overnight oximetry performed, although not all of these subjects could perform all of the measures of respiratory muscle strength. Correlation analysis of the mean nocturnal oxygen saturation levels to SNIF was r = 0.4, p = 0.001; for MIF, r = 0.322, p = 0.056; and for FVC, r = 0.39, p = 0.003 (Figure 3).
Threshold Levels of FVC, MIF, and SNIF to Predict Nocturnal Hypoxemia and Death
To evaluate the clinical relevance of the noninvasive tests for predicting mortality, we used threshold levels for each test. The association between FVC, MIF, and SNIF categories with time spent at less than 90% oxygen saturation at night are shown in Figure 4. Only SNIF category (above or below 40 cm H2O) was significantly associated with desaturation on this analysis (p = 0.01).
Predictors of Mortality in ALS
Each test was considered as a continuous variable, and Cox regression analysis was used to examine the hazard ratios for predicting death. Table E1 (in the online supplement) shows the results of this analysis. Age, body mass index, FVC (% predicted), SNIF, MIF, and SNEF were all independently associated with a significantly increased risk of death at any point of follow-up. In addition, the oximetry variables measured; mean O2 and proportion of night spent at less than 90% O2 saturation were related to death.
A SNIF less than 40 cm H2O was associated with a hazard risk for death of 9.1 (95% CI 4eC20.8, p < 0.001) (Figure E2). The 25th percentile for mortality when the SNIF fell to less than 40 cm H2O was 3.46 ± 0.1 months (95% CI 2.51eC5.52) and the median (50%) mortality was 6 ± 0.3 months (95% CI 2.51eC8.45). When the eight patients who used noninvasive ventilation were excluded from the data set, there was still no effect on the major result of the study (i.e., that SNIF < 40 cm H2O was a positive predictor of death; data not shown). Figure 5 shows the Kaplan-Meier survival curve for all subjects separated into change of SNIF of 10 cm H2O. The hazard ratio for death when the SNIF was less than 30 cm H2O compared with those with a SNIF more than 30 cm H2O was 5.9 (95% CI 3eC12). When FVC fell to less than 50%, the hazard ratio for death was 5.66 (95% CI 2.73eC11.73, p < 0.001).
Among patients with a SNIF less than 40 cm H2O, 66% had an FVC greater than 50%; in this group, the hazard ratio for death was 13.6 (95% CI 3.1eC54.7, p < 0.001) (Table 3). We defined sensitivity as the proportion of subjects who died when their measurements were FVC less than 50% and MIF and SNIF less than 40 cm H2O 6 months before death and the specificity as the proportion of subjects who lived for more than 6 months when their test was above the cutoff value. The sensitivity of FVC less than 50% for predicting 6-month mortality was 58% with a specificity of 96%, whereas for SNIF less than 40 cm H2O, the test had a sensitivity of 97% and a specificity of 79% for death within 6 months (Table 4).
DISCUSSION
The results of this study show that SNIF correlated well with transdiaphragmatic strength and so is a good measure of respiratory muscle strength. Furthermore, at the later stages of the disease, the SNIF test could still be performed by 96% of the patients. Finally, we established that a SNIF less than 40 cm H2O is associated with a median survival of 6 months. Thus the SNIF test is a sensitive means of detecting respiratory muscle strength, it can be recorded in advanced disease, and it gives important prognostic information.
Prior studies of the natural history of ALS have identified that deteriorations in pulmonary function tests are the most important prognostic factors in survival (2, 3, 13, 14). These studies have generally used the FVC as the main measure of respiratory muscle strength. The FVC is not a particularly good measure of respiratory muscle strength because the shape of the pressureeCvolume curve means that it does not detect modest falls in muscle strength. Thus alternative measures of muscle strength with MIF and SNIF testing have been investigated in some prior studies. One retrospective study demonstrated an association between MIF and survival, but did not establish a value that gave a guide to prognosis (15). Prior prospective studies have reported that composite respiratory scores—a combination of FVC and MIF—were related to survival. However, in these studies too, a threshold value for either test which provided prognostic information was not established (3, 14).
The study population was prospectively recruited and followed over 3 years. Approximately 40% of the study subjects were women, reflecting the expected male:female ratio of 1.6:1 in an ALS population. Although the average FVC was slightly reduced at the start of the study, the direct measures of muscle strength, MIF and SNIF, were considerably reduced in most patients. At the start of this study, a device to measure SNIF was not commercially available, so we modified a hand-held mouth pressure meter to record these pressures. To establish the reliability of this new device, we undertook a series of validation studies, testing the fidelity of the recording device and establishing a range of values of SNIF in a cohort of healthy volunteers of a similar age as the ALS patients. The mean value of SNIF was 90 cm H2O, which is in agreement with prior data, suggesting that our modified device was suitable for this study (16). In addition, we related SNIF, MIF, and FVC to transdiaphragmatic muscle strength in a subgroup of patients with ALS. In these studies, it was shown that both MIF and SNIF correlated better with transdiaphragmatic pressure than with FVC. This confirms the results of prior studies indicating that MIF and SNIF are better methods of recording respiratory muscle strength than FVC. We had noticed, clinically, that weakness of the bulbar muscle often prevents patients from making a complete seal around a mouth device; consequently, FVC and MIF may not be recorded accurately at a time when the information is most relevant. In the current study, 4 patients (4%) could not perform SNIF at the last visit, whereas 14 (14%) and 19 (19%) could not perform the FVC and MIF, respectively. Thus, although MIF and SNIF are better tests of respiratory muscle strength than FVC, more patients with advanced disease can perform the SNIF test, making this the preferred test to measure respiratory muscle strength in ALS.
The primary purpose of this study was to establish if a SNIF value could be established that predicted mortality in patients with ALS. We categorized SNIF measurements into units of 10 cm H2O below 50 cm H2O and related these to mortality. A SNIF less than 50 cm H2O had a relatively good prognosis, whereas a SNIF less than 40 cm H2O had a median survival of less than 6 months and a 1-year survival rate less than 25%. When the SNIF fell below 30 cm H2O, the median survival was 3 months; few patients with such a measure survived more than 6 months.
Although the primary end point of the study was mortality, we also related the tests of lung function to nocturnal oxygen saturation. Nocturnal hypoxia is a feature of hypoventilation in people with otherwise healthy lungs and is seen before daytime hypoxia and so reflects early respiratory muscle weakness. A SNIF less than 40 cm H2O was the only test that was significantly correlated with nocturnal hypoxia. Thus a fall in SNIF to this level may be an appropriate time to consider noninvasive ventilation. Studies in our unit are in progress to establish if initiation of nocturnal ventilatory support leads to increased survival when the SNIF falls to this level.
Noninvasive ventilation may prolong life and improve quality of life in patients with ALS (17eC19). However, there is no consensus on when to initiate this treatment (20). The timing of instituting ventilatory support and discussion of the overall prognosis are important issues for patients. The prognosis, the recommended timing of noninvasive ventilation, and assessment of new treatments in ALS all use FVC% as the standard measure of respiratory performance (21, 22). The results of this study highlight the limitations of using the FVC in ALS. First, it cannot be obtained in about 20% of subjects at the later stages of the disease, and second, it is not sensitive to important changes in respiratory muscle strength because 75% of subjects with important levels of muscle weakness (SNIF < 40 cmH2O) still had an FVC greater than 50%. This is important because, in this group, the hazard ratio for death within 6 months was 13.6. In summary, the SNIF is a valuable tool to monitor respiratory muscle strength in patients with ALS.
Acknowledgments
The authors are grateful to the patients who participated in this study.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
These authors contributed equally to this manuscript.
REFERENCES
Kreitzer S, Saunders NA, Tyler R, Ingram RH Jr. Respiratory muscle function in amyotrophic lateral sclerosis. Am Rev Respir Dis 1978;117:437eC447.
Fallat RJ, Jewitt B, Bass M, Kamm B, Norris FH Jr. Spirometry in amyotrophic lateral sclerosis. Arch Neurol 1979;36:74eC80.
Stambler N, Charatan M, Cedarbaum JM. Prognostic indicators of survival in ALS. ALS CNTF Treatment Study Group. Neurology 1998;50:66eC72.
Fitting JW, Paillex R, Hirt L, Aebischer P, Schluep M. Sniff nasal pressure: a sensitive respiratory test to assess progression of amyotrophic lateral sclerosis. Ann Neurol 1999;46:887eC893.
Miller RG, Rosenberg JA, Gelinas DF, Mitsumoto H, Newman D, Sufit R, Borasio GD, Bradley WG, Bromberg MB, Brooks BR, et al. Practice parameter: the care of the patient with amyotrophic lateral sclerosis (an evidence-based review). Muscle Nerve 1999;22:1104eC1118.
Lyall RA, Donaldson N, Polkey MI, Leigh PN, Moxham J. Respiratory muscle strength and ventilatory failure in amyotrophic lateral sclerosis. Brain 2001;124:2000eC2013.
Stefanutti D, Benoist MR, Scheinmann P, Chaussain M, Fitting JW. Usefulness of sniff nasal pressure in patients with neuromuscular or skeletal disorders. Am J Respir Crit Care Med 2000;162:1507eC1511.
Morgan RK, McNally S, Hardiman O, Costello RW. SNIPmax as a predictor of mortality in amyotrophic lateral sclerosis. Am J Respir Crit Care Med 2004;169:A440.
Bradley WG. Overview of motor neuron disease: classification and nomenclature. Clin Neurosci 1995;3:323eC326.
Lin LI-K. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989;45:255eC268.
Lin LI-K. A note on the concordance correlation coefficient. Biometrics 2000;56:324eC325.
Polkey MI, Green M, Moxham J. Measurement of respiratory muscle strength. Thorax 1995;50:1131eC1135.
Haverkamp LJ, Appel V, Appel SH. Natural history of amyotrophic lateral sclerosis in a database population: validation of a scoring system and a model for survival prediction. Brain 1995;118:707eC719.
Ringel SP, Murphy JR, Alderson MK, Bryan W, England JD, Miller RG, Petajan JH, Smith SA, Roelofs RI, Ziter F, et al. The natural history of amyotrophic lateral sclerosis. Neurology 1993;43:1316eC1322.
Gay PC, Westbrook PR, Daube JR, Litchy WJ, Windebank AJ, Iverson R. Effects of alterations in pulmonary function and sleep variables on survival in patients with amyotrophic lateral sclerosis. Mayo Clin Proc 1991;66:686eC694.
Uldry C, Fitting JW. Maximal values of sniff nasal inspiratory pressure in healthy subjects. Thorax 1995;50:371eC375.
Newsom-Davis IC, Lyall RA, Leigh PN, Moxham J, Goldstein LH. The effect of non-invasive positive pressure ventilation (NIPPV) on cognitive function in amyotrophic lateral sclerosis (ALS): a prospective study. J Neurol Neurosurg Psychiatry 2001;71:482eC487.
Aboussouan LS, Khan SU, Meeker DP, Stelmach K, Mitsumoto H. Effect of noninvasive positive-pressure ventilation on survival in amyotrophic lateral sclerosis. Ann Intern Med 1997;127:450eC453.
Pinto AC, Evangelista T, Carvalho M, Alves MA, Sales Luis ML. Respiratory assistance with a non-invasive ventilator (BiPAP) in MND/ALS patients: survival rates in a controlled trial. J Neurol Sci 1995;129:19eC26.
Lechtzin N, Rothstein J, Clawson L, Diette GB, Wiener CM. Amyotrophic lateral sclerosis: evaluation and treatment of respiratory impairment. Amyotroph Lateral Scler Other Motor Neuron Disord 2002;3:5eC13.
Cudkowicz ME, Shefner JM, Schoenfeld DA, Brown Jr RH Jr, Johnson H, Qureshi M, Jacobs M, Rothstein JD, Appel SH, Pascuzzi RM, et al. A randomized, placebo-controlled trial of topiramate in amyotrophic lateral sclerosis. Neurology 2003;61:456eC464.
Groeneveld GJ, Veldink JH, van der Tweel I, Kalmijn S, Beijer C, de Visser M, Wokke JH, Franssen H, van den Berg LH. A randomized sequential trial of creatine in amyotrophic lateral sclerosis. Ann Neurol 2003;53:437eC445.(Ross K. Morgan, Stephen M)