The important findings of this study are that, in this large sample of subjects with COPD selected for having an increased risk of experiencing an AECOPD within one year, only 1.9% had MBL concentrations below the normal range reported by the manufacturer of the assay, and regardless of the MBL concentration used to define MBL deficiency, we found no association between deficiency and time to first AECOPD, rate of AECOPDs, or need for subjects to be hospitalized for AECOPDs.
Two studies concluded that MBL deficiency was associated with an increased incidence of AECOPDs and one concluded that it was not. The age, spirometry, and smoking histories of the subjects in these three studies were similar to those we evaluated, with the exception that none of the three studies selected patients who were at increased risk of experiencing acute exacerbations as we did. Yang et al30
found that 24 of 82 (29%) subjects with COPD had MBL-deficient genotypes. These subjects had lower MBL concentrations than those with the wild-type genotype (107 ng/mL, IQR 30–246, range 21–7675 versus 920 ng/mL, IQR 398–1355, range 21–2256, P
< 0.001). MBL concentrations were not presented in a fashion that allowed determination of the prevalence of MBL deficiency based on the definitions used in the literature. Forty of the 82 patients (49%) had one or more admissions for AECOPDs during a two-year follow-up period and 18 of these (45%) had MBL-deficient genotypes. Forty-two patients had no AECOPDs and only six (14%) of these had MBL-deficient genotypes (odds ratio 4.9 95% CI 1.7–14.4, P
= 0.0037). We did not determine MBL genotypes but genotypes do not accurately predict MBL concentrations (see below, and note that at least one of the subjects with an MBL-deficient genotype included in the study by Yang et al had an MBL concentration of 7675 ng/mL and one with the wild-type genotype had an MBL concentration of 21 ng/mL).
Lin et al32
studied 215 Chinese subjects with COPD. MBL concentrations were again not presented in a fashion that allowed determination of the prevalence of MBL deficiency by literature-based definitions, but the median MBL concentration they reported of 777 ng/mL was comparable with the median of 918 ng/mL that we observed. In 96 of their subjects who had three or more AECOPDs per year over a 3-year period, the mean MBL concentration was 534 ± 932 ng/mL compared with 1127 ± 1469 ng/mL in subjects with two or fewer exacerbations (P
= 0.027). Twelve of the 96 more frequent exacerbators (12.5%) had MBL genotypes associated with MBL deficiencies compared with only 5/119 (4.2%) of the less frequent exacerbators (P
= 0.025). Patients with MBL-deficient genotypes also had a higher mortality (66.7% versus 31.0%, respectively, P
However, in a cross-sectional study, Eagan et al31
found that 18.2% and 42.9% of healthy Norwegian subjects had MBL concentrations ≤ 100 ng/mL and ≤500 ng/mL, respectively, compared with 22.2% and 49.6% of subjects with COPD (P
= 0.23 and P
= 0.10, respectively). No association between MBL concentration and a history of AECOPDs was observed, but no information was provided with respect to how AECOPDs were defined. Eagan et al31
also noted that subjects with GOLD stage 3 disease had a higher prevalence of MBL deficiency (defined as ≤100 ng/mL). We found no demographic or COPD severity indicators that were more or less common in subjects with MBL deficiency, regardless of how deficiency was defined ().
The strengths of our study include the large sample size, the multicenter design, and the prospective ascertainment of AECOPDs using an event-based (ie, health care utilization) definition. Our study population was enriched by enrolling subjects whom we anticipated would be at increased risk of experiencing an AECOPD within the one-year follow-up period, based on previous predictors identified by Niewoehner et al.34
This should have increased the likelihood of finding a higher prevalence of MBL deficiency, regardless of how deficiency was defined, compared with the prevalence in healthy controls.
Our study has a number of limitations. First, we only measured MBL concentrations on one occasion. Several groups have demonstrated that MBL is an acute phase reactant.41
However, in clinically stable patients, MBL concentrations are constant over time,42
and our patients had to be free of AECOPDs for at least 4 weeks before meeting inclusion criteria. In addition, even during acute phase responses, MBL concentrations in patients with low concentrations only increase by 1.5–4.3-fold and do not reach normal values.45
Accordingly, we suggest that there is a low likelihood of our data being confounded by spuriously elevated concentrations of MBL resulting in an underestimate of the prevalence of MBL deficiency.
We did not confirm by genotyping that low MBL concentrations were the result of variant alleles. Approximately 30% and 4%–8% of the normal population have heterozygous or homozygous genetic mutations, respectively, associated with low MBL concentrations.35
Although MBL concentrations are well correlated with genotypes, (eg, MBL concentrations < 50 ng/mL are 100% sensitive and 83% specific for variant exon-1 polymorphisms43
) subjects with wild-type MBL genes may still have low concentrations of MBL,35
and MBL concentrations may vary as much as 10-fold in patients with identical genotypes for all of the MBL variants described to date.38
Accordingly, assessing relationships between MBL concentrations and AECOPDs is likely to be a more sensitive approach than genotyping.21
Similarly, we did not assess MBL function using a complement deposition assay. Bouwman et al47
suggested that while MBL function was a better way of defining MBL deficiency than determining MBL concentrations, assessing MBL concentrations was “most appropriate” for defining MBL deficiency in studies seeking associations with infections. Differences between MBL binding and complement activation may vary depending on the method of assessment,49
and MBL concentrations from 500 ng/mL to 1000 ng/mL are associated with decreases in function by up to 90%.9
While we saw no association between MBL deficiency and AECOPDs, MBL deficiency could still be associated with AECOPDs that result from infection with one or more specific pathogens if the frequency with which these specific pathogens affected our subjects was too low to discern the association. However, we think this possibility is unlikely, because the two studies documenting the association between MBL deficiency and AECOPDs found an association with AECOPDs that were not otherwise defined by cause or potential infecting organism.30
In addition, Lin et al32
found no difference in the distribution of pathogens in those subjects with AECOPD and without MBL-deficient genotypes.
Because we only found 20 subjects (1.9%) with MBL concentrations below the lower range of normal reported by the manufacturer of the assay, we cannot exclude the possibility that very low concentrations of MBL have an association with AECOPDs. However, the infrequency of this finding implies that even if this association were found, it would pertain only to a small minority of patients suffering AECOPDs.
In conclusion, we found a very low prevalence of MBL deficiency in subjects who were at increased risk for experiencing AECOPDs and no association between MBL deficiency and time to first acute exacerbation, frequency of acute exacerbations, or percent of subjects requiring hospitalization for acute exacerbations in subjects with MBL deficiency regardless of how deficiency was defined. Accordingly, our data do not support the idea that MBL might be a therapeutic target to reduce the incidence of AECOPDs. Rather, they imply that, while COPD is an inflammatory disorder with systemic manifestations, the fundamental pathophysiology of COPD differs from conditions in which MBL deficiency seems to be a clear risk factor (ie, childhood pneumonia, rheumatoid arthritis, systemic lupus).