Pulmonary function in chronic spinal cord injury: A cross-sectional survey of 222 Southern California adult outpatients☆☆☆★★★♢♢♢

Article Outline

• Abstract
• Methods
  • Study population and test procedures
  • Data analysis
• Results
  • Overview: Function loss in relation to injury characteristics and smoking history
  • Function in persons with high tetraplegia (C5 and above)
  • Function in persons with low tetraplegia (C6-C8)
  • Function in persons with high paraplegia (T1-T7)
  • Function in persons with low paraplegia (T8 and below)
• Discussion
  • Effects of level and completeness of SCI
  • Effects of duration
  • Effects of smoking
  • Effects of position
• Conclusions
• References
• Copyright

Abstract 

Linn WS, Adkins RH, Gong H Jr, Waters RL. Pulmonary function in chronic spinal cord injury: a cross-sectional survey of a large southern California outpatient population. Arch Phys Med Rehabil 2000;81:757-63. Objectives: To evaluate risk factors for respiratory morbidity in chronic spinal cord injury (SCI). Setting: Model SCI care system based at an urban public rehabilitation medical center. Design: Case series with evaluation of pulmonary function by conventional spirometric testing. Participants: Two hundred twenty-two adults with SCI of more than 1-year duration who were not chronically dependent on mechanical ventilation, including 98 with tetraplegia (62 with complete and 26 with incomplete motor lesions) and 124 with paraplegia (87 with complete and 37 with incomplete motor lesions). Main Outcome Measures: Forced vital capacity (FVC), forced expired volume in 1 second (FEV₁), and peak expiratory flow rate (PEFR), all measured in the supine and erect seated positions and compared with predicted normal values for industrial workers. Results: FVC and FEV₁ were normal in persons with low-level paraplegia who had never smoked, but both decreased similarly with rising SCI level, more markedly in those with tetraplegia. PEFR decreased with rising SCI level. Incomplete lesions mitigated function loss in those with tetraplegia. In middle-aged individuals with tetraplegia, longer duration of injury was associated with greater function loss, independent of age. Current smokers showed excess function loss, except for those with high tetraplegia. Most people with complete tetraplegia showed FVC and FEV₁ increases in the supine position relative to the erect position. Conclusions: Pulmonary function is compromised by most lesions of the spinal cord, even in those with paraplegia, and is affected relative to the level of lesion. Efforts to help SCI patients minimize respiratory complications—in particular, assistance in smoking cessation—should be given high priority. © 2000 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation

Keywords:  Paraplegia, Tetraplegia, Spirometry, Smoking, Rehabilitation

PEOPLE WITH SPINAL cord injury (SCI) are at increased risk of chronic respiratory symptoms, added disability, and early death from respiratory complications.¹, ², ³, ⁴ Therefore, they require continuing careful evaluation and management of their respiratory status. Respiratory muscle paralysis both restricts maximum inflation of the lungs and impairs the ability to cough, leading to increased risk of atelectasis and retained mucus secretions.⁵, ⁶ Obstructive pulmonary dysfunction is also of concern, not only because airways may collapse or be clogged by mucus, but also because they may be especially susceptible to constriction. Bronchial hyper-reactivity is common in individuals with higher-level SCI, possibly as a result of interruption of the lungs' sympathetic innervation⁷, ⁸ and/or insufficient stretching of airway smooth muscle because of impairment of inspiratory forces.⁹

Satisfactory management requires understanding of the degree of pulmonary function loss attributable to any given level of SCI and of other accompanying risk factors that may further reduce function. Most previous investigations of these issues have been limited by relatively small numbers of patients available for study. Almenoff and colleagues¹⁰ presented the largest previous SCI/pulmonary function investigation, studying 165 male military veteran outpatients from metropolitan New York, and reviewing earlier findings. Using linear regression and analysis of variance (ANOVA), they found statistically significant correlations of SCI level with forced vital capacity (FVC), forced expired volume in 1 second (FEV₁), and peak expiratory flow rate (PEFR). Incomplete lesions mitigated dysfunction (relative to complete lesions) in tetraplegia but not in paraplegia. Current smokers with lesions at C5 and below showed clinically and statistically significant excess decrements in FEV₁, but non–ventilator-dependent individuals with C2-C4 lesions did not show a smoking-related deficit.

We used analyses similar to those of Almenoff's group¹⁰ as well as nonlinear regressions to investigate factors influencing pulmonary function in a large population with chronic SCI that was demographically different from Almenoff's. We obtained data from 187 men and 35 women outpatients at a metropolitan Los Angeles public rehabilitation hospital, a group that was predominantly Hispanic (Mexican American) but included whites (not Hispanic) and blacks (African Americans). The group had a broad age range but was younger on average than Almenoff's population. We evaluated the effects of SCI level, completeness of lesion, smoking history, duration of injury, and body position (erect vs supine) on FVC (a measure of restrictive dysfunction), PEFR (a measure of cough effectiveness), and FEV₁ (a measure of obstructive dysfunction when expressed as the ratio FEV₁/FVC).

Back to Article Outline

Methods 

Study population and test procedures 

The potential study population consisted of 235 consecutive adult outpatients with SCI of more than 1-year duration who were not chronically dependent on mechanical ventilation and were undergoing a comprehensive physical examination at Rancho Los Amigos National Rehabilitation Center. Six potential subjects were excluded because predicted values were not available for their ethnic groups (see below), and seven because of missing predictor data, leaving 222 analyzable subjects. Table 1 presents statistics on age, size, and injury characteristics at the time of pulmonary function testing. Height was measured as supine length; weight was measured with the subject in a chair of known weight. From questionnaire responses, subjects were classified as never-smokers (fewer than 100 cigarettes lifetime, n = 111), former smokers (no smoking in past 6 months, n = 56), and current smokers (actively smoking or quit for less than 6 months, n = 55). Lifetime smoking exposure was quantified in pack-years (1 pack-year = 1 pack of cigarettes per day for 1 year) based on self-reports. The level of injury was defined as the lowest normal motor segment, determined by neurologic examination following American Spinal Injury Association guidelines.¹¹ Subjects were categorized as having high tetraplegia (injury level C2-C5, impairment of diaphragm function expected), low tetraplegia (C6-C8, normal diaphragm function expected), high paraplegia (T1-T7, sympathetic impairment expected), or low paraplegia (T8-L5, normal sympathetic function expected).

Table 1: Characteristics of 222 outpatients studied (mean and range, by sex and ethnicity)

Abbreviations: F, female; M, male.

All sex/ethnic subgroups were diverse in size, age, and injury characteristics (table 1). Women's age and duration of SCI averaged significantly higher than men's. Hispanics averaged significantly younger and had shorter durations than non-Hispanics. However, distributions of SCI levels were not significantly different between ethnic groups, nor between sexes.

Pulmonary function was measured according to American Thoracic Society guidelines¹² using a computer-interfaced automated testing and analysis system, either with a pneumotachograph sensor (Vmax System^a) or with a rolling-seal volumetric spirometer (System 7200^a). Instruments were calibrated daily according to the manufacturer's procedure, using a volumetric syringe. Each subject was tested in seated-erect and supine positions, in random order, on the same day. Supine testing was performed by lowering the wheelchair back or by transferring the subject to a flat examination table. Seated-erect measurements are reported, except where supine measurements are mentioned specifically.

Data analysis 

Analyses were performed with BMDP statistical software.^b Effects were considered statistically significant at p < .05. Function measurements were expressed as percentages of normal predicted values (FVC%, FEV₁%, and PEFR%) from a study of nonsmoking blue-collar workers by Glindmeyer and coworkers.¹³ These able-bodied control subjects were most appropriate for our population because (1) prediction equations accounted for possibly accelerated function loss at higher ages, (2) separate predictions were available for black and white subjects tested similarly, and (3) exclusion of anyone over age 65 minimized upward bias in predictions for older subjects attributable to a survivor effect. A few of our subjects were over 65, but their data appeared to fit the predictions, so they were included in analyses.

The able-bodied reference population did not include Hispanics (H. Glindmeyer, personal communication, April 1998). However, predictions for whites were considered valid for our Hispanic subjects, because their FVC% and FEV₁% were not significantly different from those of non-Hispanic whites. In addition, a recently published survey of a national population sample¹⁴ confirms white-Hispanic similarity and shows generally good agreement with the predictors we used.¹³ Subjects of Asian or “other” ethnicities were too few for meaningful statistical comparisons with larger groups. Their data appeared not to match either white or black predictions, so they were excluded from analysis. To validate predictions for white (including Hispanic) and black subjects, preliminary regression analyses were performed to relate FVC% or FEV₁% with age, sex, and ethnicity, including only individuals with low paraplegia and presumably minimal SCI-related impairment. Additional preliminary analyses were performed for all subjects, including analysis of other significant predictors (see below). Age, sex, ethnicity, and body mass index (weight/height²) effects were not significant; thus, the prediction method was considered valid.

To assess the influences of SCI level and other risk factors, linear regressions were performed with FVC%, FEV₁%, or PEFR% as the dependent variable. Independent variables were level of injury (numbered lowest to highest, ie, S5 = 1, C1 = 30), duration of injury (age at testing minus age at SCI, in years), incompleteness (0 = complete, 1 = incomplete), and smoking pack-years. Smoking-related declines in pulmonary function tend to be partially reversed after smoking cessation,¹⁵ so we employed one pack-year variable for current smokers (equal to 0 for never and former smokers) and another for former smokers (equal to 0 for never and current smokers). Linear regressions were performed separately for the four injury-level categories, because results differed between them. For FVC and FEV₁, linear predictions for the four categories did not fully capture the inherently nonlinear physiologic relationship to SCI level (see the Results section). As an alternative, we regressed function measures against level, level squared, duration, and the two pack-year variables for all subjects with complete lesions to derive a single prediction equation for all SCI levels. We computed the difference between this equation's prediction and the observed function value for each subject with an incomplete lesion, then regressed those differences against level and level squared. This second regression estimated the degree to which an incomplete motor lesion mitigated function loss at any SCI level. To assess the effect of body position, we expressed supine FVC%, FEV₁%, and PEFR% as percentages of erect measurements, then investigated the influences of the various predictors by regression and ANOVA. Alternative analyses of absolute rather than percentage changes (not reported) gave similar results.

Back to Article Outline

Results 

Overview: Function loss in relation to injury characteristics and smoking history 

For each SCI level category, table 2 shows descriptive statistics, and table 3 shows results of linear regressions relating pulmonary function to injury level, completeness, duration, and smoking history. Due to limitations of the data and the analytical model, linear regressions sometimes gave inconsistent predictions at adjacent SCI levels in different categories (better function at higher levels). The alternative analyses of complete-lesion subjects, with linear and quadratic terms for level, yielded the following prediction equations: FVC% = 93.3 + 2.21L − 0.143L² − 0.277D − 0.93P_C − 0.04P_F FEV₁% = 90.9 + 2.57L − 0.150L² − 0.181D − 0.92P_C − 0.0001P_F where L = level (S5 = 1, C1 = 30), D = duration (yrs), P_C = pack-years in current smokers, and P_F = pack-years in former smokers. These equations appeared to fit the individual data equally well at high and low levels. Decreases with rising SCI level and with increasing pack-years in current smokers were statistically significant (p < .05). The decrease with increasing duration of injury was significant for FVC%, but not for FEV₁%. The effect of pack-years in former smokers was not significant and close to zero. The proportion of variance explainable by the predictor variables was 62% for FVC and 57% for FEV₁. Figure 1 shows expected FVC% at each level, based on the above equation, for subjects with complete lesions of 15 years' duration who have never smoked (filled circles) or who currently smoke and have a lifetime exposure of 20 pack-years (filled triangles).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Predictions from nonlinear regression analysis: Forced vital capacity (FVC) 15 years after spinal cord injury in relation to level of injury for never-smokers with complete lesions (●), never-smokers with incomplete lesions (○), and current smokers with 20 pack-years lifetime exposure (▴).

Open circles indicate expected FVC% for never-smokers with incomplete lesions of 15 years' duration, based on the above prediction equation plus the second regression analysis of incomplete lesions described in the Methods section.

Table 2: Pulmonary function (% of predicted value) in never-smokers with complete lesions

Abbreviations: FEV₁, forced expired volume in 1 second; FVC, forced vital capacity; PEPR, peak expiratory flow rate.

Table 3: Influences of injury level, completeness, duration, and smoking history on pulmonary function (% of predicted value), estimated by linear regression

Abbreviations: FEV₁, forced expired volume in 1 second; FVC, forced vital capacity; PEFR, peak expiratory flow rate.

For PEFR%, the level-squared term explained very little variance: the loss of PEFR with increasing level was essentially linear. The PEFR% decreased by 2.7 percentage points per each level rise in neurologic level of injury (p < .005), roughly the same as found previously by Wang's group.⁶ The PEFR% also decreased by 0.9 percentage points for each additional pack-year in current smokers (p < .05). Effects of duration and of smoking in former smokers were not significant. Only 44% of variance in PEFR% was explainable by the predictor variables; this was not surprising, as PEFR is less predictable from size and age than FVC or FEV₁.¹³ Incomplete lesions were associated with significant increases in PEFR% at higher levels of injury, but statistical findings were anomalous in that people with incomplete low paraplegia appeared to have generally lower PEFR% than those with complete low paraplegia (Table 3).

Regression analyses of the FEV₁/FVC ratio showed no significant relationships to SCI level, duration, or smoking history. There was a small significant decrease (averaging about 3% for the entire subject population) with incomplete lesions relative to complete lesions. That is, incomplete lesions mitigated FEV₁ loss slightly less than FVC loss. This effect tended to be larger in tetraplegia than in paraplegia, but the difference was not significant.

Function in persons with high tetraplegia (C5 and above) 

The typical function loss associated with complete motor lesions at C5 and above was roughly one half for FVC or FEV₁ and more than one half for PEFR (table 2). Virtually all subjects with complete high tetraplegia fell below statistical lower limits of normal results for the able-bodied reference population,¹³ which varied with age and size but averaged near 79 for FVC%, 77 for FEV₁%, and 75 for PEFR%. Incomplete lesions mitigated losses in all three function indices to a clinically and statistically significant extent, such that some individuals with incomplete lesions fell within normal limits.

Current smokers with high tetraplegia showed increases in function with increasing pack-years, which reached statistical significance for FVC% and PEFR%, whereas former smokers showed no significant effect (table 3). The effect of injury level was essentially zero for PEFR%, nonsignificantly negative for FEV₁%, and significant (but less negative than in low tetraplegia) for FVC%.

Persons with high tetraplegia showed a significant decrease in FVC% with increasing duration of injury, estimated at .42 percentage points per year (table 3). Those with low tetraplegia showed about the same estimated loss with increasing duration, although it was not statistically significant. Thus, individuals with tetraplegia of long duration appeared to have lost more vital capacity than expected from their level of injury and from the usual effects of aging in able-bodied persons. This predicted excess loss would be clinically important by middle age for subjects injured in their teens or early twenties. Because most injuries occurred in that age range, duration of injury and chronological age were highly correlated (r = .66 for all subjects), making it difficult to differentiate the effects of these parameters on pulmonary function. To address that problem, we restricted the analysis to persons aged 35 to 55yrs with tetraplegia, for whom age and duration were not correlated (r = .002). After allowing for SCI level, incompleteness, and pack-years, effects of age were nonsignificantly positive, but effects of duration were negative, statistically significant, and clinically significant, indicating about a 15-percentage-point loss 20 years after injury (table 4). These results are consistent with an effect of duration: among middle-aged people with tetraplegia, those injured earlier had consistently worse function than those injured later.

Table 4: Comparative influences of age and duration of injury in subjects aged over 35 but under 55 years, estimated by linear regression

Abbreviations: FEV₁, forced expired volume in 1 second; FVC, forced vital capacity.

For people with complete high tetraplegia, mean FVC, FEV₁, and PEFR all showed modest but statistically significant increases in the supine position relative to the erect position (table 5). Individuals varied greatly. Some showed marked function improvements when supine while others showed decrements. For those with incomplete lesions, function did not change significantly between erect and supine positions.

Table 5: Pulmonary function in supine position, as a percentage of function in erect position, in subjects with complete lesions

Abbreviations: CI, confidence interval; FEV₁, forced expired volume in 1 second; FVC, forced vital capacity; PEFR, peak expiratory flow rate.

Function in persons with low tetraplegia (C6-C8) 

The typical function loss in people with C6-C8 complete motor lesions who had never smoked was about one third for FVC or FEV₁, and nearly one half for PEFR (table 2). Most individuals fell below statistical lower limits of normal. The typical additional loss at each successively higher neurologic level of injury was statistically significant and large—nearly 9 percentage points—for FVC% and FEV₁%. In contrast, PEFR% showed no dependence on injury level within the low tetraplegia range. In low as in high tetraplegia, incomplete lesions appreciably mitigated function loss (table 3) so that some individuals fell within the normal range.

Former smokers with low tetraplegia showed no appreciable function differences from never-smokers. Current smokers showed statistically and clinically significant decrements in FVC% and FEV₁%—about one-half-percentage point loss per pack-year—but no appreciable effect on PEFR% (table 3). The distribution of pack-year values was skewed, with few subjects high enough that clinically significant smoking effects would be expected, judging from the able-bodied population (see Discussion section). To test whether the few heavy/long-term smokers disproportionately influenced the overall results, data were reanalyzed after excluding current smokers with more than 10 pack-years. The relationship between FVC or FEV₁ and pack-years remained significant.

In people with complete low tetraplegia, mean FVC, FEV₁, and PEFR all increased modestly in the supine position relative to the erect position (table 5). Only the FVC change was statistically significant. Individuals' changes varied from markedly positive to markedly negative. In those with incomplete low tetraplegia, function did not change significantly between erect and supine positions.

Function in persons with high paraplegia (T1-T7) 

In subjects with T1-T7 complete motor lesions, function ranged from well below to well within normal limits, averaging slightly above the lower limits of normal for FVC and FEV₁ and slightly below the lower limits of normal for PEFR (table 2). Incomplete lesions were associated with function improvements averaging a few percent, which was not statistically significant (table 3). The estimated additional function loss with each successive rise in neurologic level of injury was between 1 and 2 percentage points for FVC% and FEV₁%, smaller for PEFR%, but not significant for all three function variables (table 3).

Former smokers with high paraplegia showed no meaningful differences from never-smokers in FVC% and FEV₁%, and slightly but significantly higher PEFR% (table 3). Current smokers showed large decrements in FVC% and FEV₁%—about 1 percentage point per pack-year (table 3). However, only the FVC effect was statistically significant, and it lost significance when current smokers with more than 10 pack-years were excluded from the analysis. PEFR% decreased nonsignificantly by .65 percentage points per pack-year.

For all subjects with high paraplegia (and for all those with low paraplegia as well), the estimated effects of injury duration on function were small and not significant (table 3). A further analysis of duration effects versus age effects was performed for subjects with paraplegia aged 35 to 55. For them, age and duration were correlated (r = .44), but less so than for the entire subject population. The estimated effect of injury duration on FVC% and FEV₁% was negative, but smaller than in people of the same age range with tetraplegia and not significant (table 4). The effect of age was nonsignificantly positive.

In people with complete high paraplegia, function was not significantly different between the erect and supine positions for FVC but was significantly lower in the supine position for FEV₁ and PEFR (table 5). Those with incomplete high paraplegia showed no significant differences between positions. As with other groups, individuals varied markedly in their responses to position change.

Function in persons with low paraplegia (T8 and below) 

Subjects with complete low paraplegia averaged only a few percent below predicted values for FVC and FEV₁, and essentially all of them fell within the normal range for the reference population. However, their mean PEFR was significantly below 100% of predicted, and some individuals fell below the normal range (table 2). In the low paraplegia range, incomplete lesions had essentially no effect on FVC or FEV₁ but were associated with a 10-percentage point decrease in PEFR% (table 3); however, that decrease was not statistically significant.

Current smokers with low paraplegia showed large significant decrements in FVC% and FEV₁% per pack-year, and modestly smaller, nonsignificant decrements in PEFR%, similar to the smoking effects in high paraplegia. Former smokers with low paraplegia showed essentially no effects (table 3). The negative relationship of FVC% with pack-years remained significant when subjects with more than 10 pack-years were excluded, although the FEV₁% relationship became nonsignificant.

The effect of injury duration on people with low paraplegia was not significant, as described in the preceding section and , . In the supine position, subjects with complete low paraplegia showed nonsignificant small decreases in FVC and PEFR relative to the erect position and a significant small decrease in FEV₁ (table 5). For subjects with incomplete low paraplegia, like those with higher-level incomplete lesions, position change caused no significant effects.

Back to Article Outline

Discussion 

Effects of level and completeness of SCI 

Our results generally corroborate the findings of previous cross-sectional SCI/pulmonary function surveys¹⁰ but differ in some details. Individuals with incomplete tetraplegia in our study, regardless of injury level, showed statistically and clinically significant preservation of function relative to those with complete tetraplegia, whereas Almenoff's group¹⁰ found significant improvement only with incomplete lesions above C5. Both our study and Almenoff's indicate that completeness of lesion has relatively little effect on pulmonary function loss in paraplegia.

Although all our nonsmokers with low paraplegia had clinically normal FVC and FEV₁, like those with high paraplegia they showed a statistical dependence of function on level of injury. This would be expected on physiologic grounds, because lesions as low as L4 disable some abdominal respiratory muscles. Thus, low paraplegia, at least from L4 upward, must be considered a risk factor for pulmonary disability: it imposes a small decrement that, if added to decrements from other risk factors, increases the chance that pulmonary function will eventually become clinically abnormal. Obviously, the risk increases with higher injury levels. Longitudinal studies are needed to determine at what point the risk assumes practical importance, in terms of injury level and duration.

The generally similar statistical results for FVC and FEV₁ and the lack of significant variation in the FEV₁/FVC ratio (except for the slight decrease with incomplete lesions) suggest that airway obstruction is clinically insignificant in our subject population in comparison with restrictive dysfunction. However, this testing may not detect early subclinical airway obstruction. The risk of future obstructive disease must be taken seriously, particularly for individuals who continue to smoke.

Peak flow results differed from FVC or FEV₁ results in that (1) PEFR averaged well below 100% of the predicted value, even in low paraplegia, and (2) PEFR% losses were approximately proportionate to SCI level throughout the range and were not disproportionately large in tetraplegia. Lower-than-predicted PEFR might be an artifact of spirometer differences between our study and the reference population study,¹³ given that PEFR measurements are more device-dependent and more difficult to standardize than FVC or FEV₁ measurements.¹² If the apparent diminution in PEFR is real, it suggests that even persons with low paraplegia and normal FVC and FEV₁ may experience some loss in cough effectiveness. (Although large PEFR losses decrease FEV₁, smaller PEFR losses may not.¹⁶) Larger, more detailed studies will be required to determine whether this phenomenon is real and clinically important. The linear relationship of PEFR% to SCI level appears consistent with findings by Wang and coworkers⁶ in a demographically different population, although their data were analyzed differently, and no comparisons with FVC or FEV₁ were available. Our findings contrast with Wang's in that we found significant PEFR improvements in incomplete tetraplegia relative to complete tetraplegia. Differences in lesion classifications and analysis techniques, as well as chance differences in subject selection, might explain this contrast.

Effects of duration 

Decreasing pulmonary function with increasing duration of injury was suggested in our entire subject population and was clearly present in those of middle age with tetraplegia. This finding is subject to alternative interpretations with different clinical implications. One possibility is that often-repeated episodes of respiratory infection cause cumulative irreversible pulmonary function decrements. Another possibility is that infection-related losses are infrequent but large, and the likelihood of having experienced one is proportional to the time elapsed since injury. Either of these would argue for more rigorous efforts to prevent and treat respiratory illnesses. Still another possibility is that existing respiratory management usually maintains stable pulmonary function in the chronic phase of SCI despite infections, and that more recently injured patients have preserved more pulmonary function because of improved management in the acute phase. Nevertheless, as with any cross-sectional study of duration of SCI, this study is subject to survivability effects. In this regard, it would be expected that the long-term survivors have maintained relatively healthful conditions, whereas health status in those with shorter durations would be more variable and perhaps poorer on average. From this latter perspective, the decline related to duration may be of even greater consequence. Longitudinal studies are needed to distinguish among these possibilities.

Effects of smoking 

Our population presented unusual problems in the analysis of smoking effects, in that: (1) typical self-estimated lifetime smoking doses in pack-years were below the range in which chronic function loss would be expected; (2) the distribution of pack-years was highly skewed, with only a few individuals reporting prolonged and/or heavy smoking; and (3) benefits usually expected from smoking cessation¹⁵ might or might not occur in people with SCI. The usual problems were also present: recall/reporting of smoking history may be inaccurate, and pack-years give a poor estimate of effective dose because of diverse individual smoking patterns. Thus, results must be interpreted with caution.

The apparent pulmonary function deficit in current smokers was striking, given their comparatively low average lifetime dose. Excluding the few individuals with high pack-years from the analysis did not change results markedly. The estimated loss was unexpectedly large, compared with previous findings in the SCI patients of Almenoff and colleagues¹⁰ or in able-bodied populations. In a cross-sectional study of office workers in the 1970s, Grimes and Hanes¹⁷ derived regression equations predicting about 9% loss in FEV₁ after 20 years at one pack per day—only about half our estimated loss. In a similar study, Azen's group¹⁸ reported an average 7% FEV₁ deficit in male smokers with higher average age and pack-years than our subjects. Grimes and Hanes found FVC decrements roughly similar to FEV₁ decrements, as we did; but Azen found no smoking-related decrement in FVC for men and small decrements, similar for FVC and FEV₁, in women. Findings in Almenoff's SCI patients somewhat resembled the findings in office workers: current smokers (modestly older than ours) averaged 77% of predicted FEV₁ and never-smokers averaged 85% (using normal parameters different from ours), but FVC was little affected by smoking.

In people with high tetraplegia, Almenoff found no significant difference between current and never-smokers, whereas we found significantly better pulmonary function in current smokers. These counterintuitive findings are explainable because high tetraplegia itself imposes a function loss sufficient to overwhelm the typical smoking-related loss. Furthermore, selection bias may have a strong effect on persons with high tetraplegia, given their broad possible range of physiologic and clinical outcomes. The more physiologically impaired individuals should be more prone to respiratory symptoms, and thus more strongly inhibited from smoking. Less impaired/less symptomatic individuals, conversely, should be more likely to smoke. Again, survivability effects in the sample may have influenced the findings. In any event, findings in people with low tetraplegia or paraplegia argue that smoking is at least as deleterious to pulmonary function for them as for able-bodied people; that should be true for high tetraplegia as well. Longitudinal studies of smokers and nonsmokers with SCI at all levels should more clearly document the long-term consequences of smoking. Even the existing data are sufficient to justify increased antismoking efforts targeted to the SCI population.

Effects of position 

We found that, on average, people with tetraplegia increased their pulmonary function measurements in the supine relative to the erect position. This is consistent with previous findings by Estenne and De Troyer¹⁹ and by Baydur and colleagues²⁰ in a subset of our population. Although the physiologic mechanism is understood,¹⁹, ²⁰ the clinical implications of this effect have not been fully explored. The increase in pulmonary function in the supine position may be substantial for some individuals whose erect pulmonary function is marginal, raising the possibility of clinically meaningful improvement of ventilation simply from lying down during episodes of dyspnea. Possible unfavorable side-effects would include increased ventilation/perfusion mismatch in the supine position and increased chance of pooled secretions leading to airway obstruction or infections. A controlled trial to explore these issues may be appropriate.

Back to Article Outline

Conclusions 

Limitations of this study include relatively small sample sizes after subdivision by SCI characteristics and other relevant factors, lack of information that might further explain pulmonary function variations (ie, respiratory health history and postbronchodilator function tests), and lack of longitudinal information needed to interpret time-dependent changes. Further studies are in progress to address these issues. Present results enhance understanding of how pulmonary function depends on the level and completeness of SCI and indicate slight but potentially important function deficits even at relatively low levels of paraplegia. Present results also suggest that in tetraplegia, duration of injury is an important determinant of pulmonary function loss, independent of aging and smoking effects. An important goal of follow-up longitudinal studies will be to determine whether this is an artifact of improving management of acute SCI or a real deterioration with time in chronic SCI. In the latter case, it will be important to identify risk factors and possible interventions. Another important follow-up goal is to clarify the effects of smoking. From the present results, people with SCI appear to be unusually susceptible to smoking-related pulmonary function loss. This “excess” smoking effect was not obvious in an earlier study¹⁰ and might be a result of chance coupled with a different distribution of smoking characteristics in our study population compared with the veteran population used previously, which also may have been affected by secondhand smoke to a greater extent. In any event, smoking must present an increased risk of long-term harm in SCI, as in any condition involving pre-existing pulmonary function impairment. All these results highlight the importance of careful long-term respiratory health care as a component of SCI management.

Back to Article Outline

References 

1. DeVivo MJ, Kartus PL, Stover SL, Rutt RD, Fine PR. Cause of death for patients with spinal cord injuries. Arch Intern Med. 1989;149:1761–1766
   • View In Article
   • MEDLINE
2. DeVivo MJ, Black KJ, Stover SL. Causes of death during the first 12 years after spinal cord injury. Arch Phys Med Rehabil. 1993;74:248–254
   • View In Article
   • MEDLINE
3. Jackson AB, Groomes TE. Incidence of respiratory complications following spinal cord injury. Arch Phys Med Rehabil. 1994;75:270–275
   • View In Article
   • MEDLINE
   • CrossRef
4. Spungen AM, Grimm DR, Lesser M, Bauman WA, Almenoff PL. Self-reported prevalence of pulmonary symptoms in subjects with spinal cord injury. Spinal Cord. 1997;35:652–657
   • View In Article
   • MEDLINE
5. Laing IS, Lammertse DP. The respiratory system in spinal cord injury. Phys Med Rehabil Clin North Am. 1992;3:725–740
   • View In Article
6. Wang AY, Jaeger RJ, Yarkony GM, Turba RM. Cough in spinal cord injured patients: the relationship between motor level and peak expiratory flow. Spinal Cord. 1997;35:299–302
   • View In Article
   • MEDLINE
7. Spungen AM, Dicpinigaitis PV, Almenoff PL, Bauman WA. Pulmonary obstruction in individuals with cervical spinal cord lesions unmasked by bronchodilator administration. Paraplegia. 1993;31:404–407
   • View In Article
   • MEDLINE
   • CrossRef
8. Singas E, Lesser M, Spungen AM, Bauman WA, Almenoff PL. Airway hyperresponsiveness to methacholine in subjects with spinal cord injury. Chest. 1996;110:911–915
   • View In Article
   • MEDLINE
   • CrossRef
9. Fein ED, Grimm DR, Lesser M, Bauman WA, Almenoff PL. Effects of ipratropium bromide on histamine-induced bronchoconstriction in subjects with cervical spinal cord injury. J Asthma. 1998;35:49–55
   • View In Article
   • MEDLINE
   • CrossRef
10. Almenoff PL, Spungen AM, Lesser M, Bauman WA. Pulmonary function survey in spinal cord injury: influences of smoking and level and completeness of injury. Lung. 1995;173:297–306
    • View In Article
    • MEDLINE
11. American Spinal Injury Association . Standards for neurological and functional classification of spinal cord injury. Chicago: American Spinal Injury Association; 1992;
    • View In Article
12. American Thoracic Society . Standardization of spirometry: 1987 update. Am Rev Respir Dis. 1987;136:1285–1298
    • View In Article
    • MEDLINE
13. Glindmeyer HW, Lefante JJ, McColloster C, Jones RN, Weill H. Blue-collar normative spirometric values for Caucasian and African-American men and women aged 18 to 65. Am J Respir Crit Care Med. 1995;151:412–422
    • View In Article
    • MEDLINE
14. Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med. 1999;159:179–187
    • View In Article
    • MEDLINE
15. Bates DV. Respiratory function in disease. 3rd ed. Philadelphia: W.B. Saunders; 1985;
    • View In Article
16. Krowka MJ, Enright PL, Rodarte JR, Hyatt RE. Effect of effort on measurement of forced expiratory volume in one second. Am Rev Respir Dis. 1987;136:829–833
    • View In Article
    • MEDLINE
17. Grimes CA, Hanes B. Influence of cigarette smoking on the spirometric evaluation of employees of a large insurance company. Am Rev Respir Dis. 1973;108:273–282
    • View In Article
    • MEDLINE
18. Azen SP, Linn WS, Hackney JD, Jones MP, Schoentgen S. A comparison of eight lung function indices in smoking and non-smoking office workers. Lung. 1977;154:213–221
    • View In Article
    • MEDLINE
    • CrossRef
19. Estenne M, De Troyer A. Mechanism of the postural dependence of vital capacity in tetraplegic subjects. Am Rev Respir Dis. 1987;135:367–371
    • View In Article
    • MEDLINE
20. Baydur A, Adkins R, Milic-Emili J. Pulmonary function in spinal cord injury: effects of injury level and posture. [abstract] Am J Respir Crit Care Med. 1998;157:A87
    • View In Article