Cigarette smoking prior to blood sampling acutely affects serum levels of the chronic obstructive pulmonary disease biomarker surfactant protein D

To the Editor,

Biomarker tests in pulmonary medicine show great promise with regard to improving patient care, yet their translation into widely used clinical tests is a slow and rather ineffective process. Very few biomarkers pass the crucial stages of the biomarker development pipeline (e.g. analytical validation, clinical validation, establishing broad clinical utility), and many efforts are thus needed to secure a good return on biomarker development investments by eventually providing health care professionals with clinically useful tests [1], [2].

Reliable analytical methods are cornerstones of biomarker testing, and the clinical usefulness of such methods depends on whether or not preanalytical variables, which may potentially affect the validity of tests results, can be controlled. In chronic obstructive pulmonary disease (COPD), there has been a distinct focus on ensuring such analytical validity [1]. For example, cigarette smoking was recently identified as a critical preanalytical variable for serum measurements of the soluble receptor for advanced glycation end products (sRAGE), a promising and (predominantly) lung-derived biomarker candidate for emphysema severity assessment in COPD [3]. Corresponding findings put previously reported associations between sRAGE and specific COPD characteristics into a different perspective given that the “acute smoking status” before blood sampling is typically not reported to be controlled in clinical biomarker studies (unlike the questionnaire-based “long-term smoking status”).

In this study, we examined the acute effects of cigarette smoking on serum levels of surfactant protein D (SPD), which represents another promising and (predominantly) lung-derived COPD biomarker candidate. This protein is present in pulmonary surfactant and is involved in the innate immune defense against various pathogens [1]. Higher SPD levels were reported for COPD patients compared with control subjects, and elevated SPD levels were found to be associated with exacerbations, emphysema progression, and mortality [1], [2]. Furthermore, previous publications revealed associations between questionnaire-based smoking status (nonsmoker vs. current smoker) and circulating SPD levels [4], [5], and we hereby aimed to explore these findings by studying the effects of cigarette smoking on SPD levels experimentally.

To this end, biobanked serum samples (stored at −80 °C for approx. 5 years) were obtained from an acute smoking study (NCT00807469 [6]), which included COPD patients, young and old individuals that have a low familial risk to develop COPD, and young individuals that have a high familial risk to develop COPD (see Table 1). In the corresponding study, serum samples were taken at baseline and 2 h after smoking three cigarettes within 1 h. Before cigarette smoking, subjects did not smoke for 2 days, which was checked by means of exhaled carbon monoxide (CO) measurements using the Micro+Smokerlyzer (a detailed overview of the CO measurements in this study to verify smoking cessation and to check smoking efficiency is provided in [7]). Blood samples were collected as described previously [8], the study was approved by the medical ethical review board of the University Medical Center Groningen (METc 2008/136), and the study adhered to the Declaration of Helsinki. In all samples, serum SPD was quantified using a validated liquid chromatography-mass spectrometry (LC-MS) method targeting the SPD protein by means of the SPD-specific peptides NEAAFLSMTDSK and SAAENAALQQLVVAK [9].

In all four study groups, similar patterns of cigarette smoke-induced SPD level changes were observed (Mann-Whitney U test; p≥0.28), thereby disqualifying SPD as susceptibility marker (based on how susceptibility was defined in the respective clinical study [6]). Data from all study groups were thus combined, revealing a statistically significant increase of serum SPD levels after cigarette smoking (one-sample t-test, p<0.0001; see Figure 1A), irrespective of the initial SPD level (see Figure 1B). Moreover, potential associations based on linear regression between the combined relative SPD level changes, as dependent variable, and the individual variables listed in Table 1, as independent variable, did not reveal any other significant association (linear regression; p≥0.21).

Figure 1:

Relative changes between SPD levels measured in serum samples that were taken 2 h after smoking three cigarettes within 1 h and samples that were taken at baseline (n=89) presented as (A) histogram and (B) Bland-Altman plot.

For preparation of these figures, data from all four study groups were combined because of the absence of statistically significant group differences (Mann-Whitney U test; p≥0.28). Figures containing data for the separate groups are included as Supplementary Figures 1 and 2.

Our study thus revealed an acute effect of cigarette smoking on serum SPD levels and substantiates the previously reported associations between questionnaire-based smoking status and circulating SPD levels [4], [5]. Reported findings should be explored in different and larger populations, and further research on the mechanistic nature of this effect is warranted. Nonetheless, it is recommended to put a tight control of cigarette smoking as a source of preanalytical variability into practice for future studies on this promising and (predominantly) lung-derived protein biomarker. This recommendation thereby supports the previously reported recommendation to standardize blood sampling conditions for SPD, which emanated from the observations that SPD exhibits some degree of circadian variation and that SPD levels are influenced by physical activity before blood sampling [10].

An important consideration with regard to the reported findings is the fact that we measured SPD levels using a validated LC-MS method, which detects SPD by means of two protein-specific peptides in the C-type lectin, ligand-binding domain of the protein [9]. This method showed adequately low bias (accuracy; within ±15%) and coefficient of variation (precision; ≤15%) values during method validation (see [9]). Matching data for in-study quality control (Supplementary Figure 3) and incurred sample reanalysis samples (see [9]) were observed during clinical sample analysis, which were all in agreement with prevailing regulatory guidelines [11], thereby supporting the analytical relevance of the observed changes. Furthermore, this method showed a very good correlation (R²=0.9; average (Bland-Altman) bias=+37%; n=32) with a commercial ELISA (BioVendor, cat. no. RD194059101), which holds an “in vitro diagnostic (IVD)” status for the European Union (see [9]). A similar correlation (R²=0.9; average [Bland-Altman] bias=−3%; n=14) was also found when comparing both methods based on the change in SPD levels due to cigarette smoking (see [9]), which argues against a method-specific artifact underlying the observed effect. At last, extensive sample stability parameters (e.g. 5× freeze-thaw stability, 27-day benchtop stability) were addressed during method validation (see [9]) to ascertain that SPD in serum is not susceptible to storage-related interferences. Extrapolating such data to the specific conditions that applied to the study samples should admittedly be done prudently, as holds true for most studies targeting biobanked samples. Nonetheless, these stability data indicate that serum SPD is a rather stable marker, at least when measured with the validated LC-MS method, thereby further supporting the plausibility of a true biological effect underlying the cigarette smoke-induced changes observed in this study.

In conclusion, cigarette smoking before blood sampling was found to induce acute changes in serum levels of SPD and should thus be considered as a critical preanalytical variable for this (predominantly) lung-derived protein. Based on these findings and similar findings for sRAGE, as reported previously, but also because of the apparent ineffectiveness of biomarker development in pulmonary medicine, we believe that we should consider controlling, or at least monitoring, a person’s smoking status before (blood) sampling for basically any lung-derived biomarker. To this regard, it may be useful to measure smoke-related compounds in exhaled breath (e.g. CO) or in the circulation to gain insight into or to verify a person’s acute smoking status.