Quantitative Analysis of Sputum Cytologic Differences Between Smokers and Nonsmokers Gary E. Swan, Ph.D., G. Berry Schumann, M.D., Tina J. Roby, M.S. CT (ASCP) (IAC), and Kent W. Sorensen
An exploratory analysis of sputum cytology was conducted to determine the extent and nature of cytomorphologic diferences between smokers and nonsmokers. Specijically, alveolar macrophages, pigmen tation of macrophages, n eu trophils, m ucus, mucous spirals, columnar cells, metaplastic cells, and dysplastic cells were analyzed and compared in 349 smokers and 93 nonsmokers. Results indicate that smokers have signijicantly higher scores on all components than do nonsmokers. Pigmentation of macrophages, alveolar macrophages, and mucous spirals were the major components discriminating between smokers and nonsmokers. A comparison of the nonsmoking sample with a diferent group of smokers (n = 466) revealed similar findings. Using the cytomorphologic parameters as the basis for classifcation, discriminant function analysis was able to classify correctly 94 percent of the nonsmokers and 91 % of the smokers, a finding confirmed in the second group of smokers. These data indicate that smokers d@er reliably from nonsmokers with respect to their cytomorphologic profile. Diagn Cytopathol 1991;7:569-575. Key Words: Sputum cytology; Cigarette smoking; Quantitative
sputum cytology Cigarette smoke is the leading known cause of primary lung cancer and chronic obstructive pulmonary disease (COPD). The field of cytopathology has made significant contributions over the years to the detection of lung lesions in their earliest stages. However, the mortality rates for both lung cancer and COPD continue to rise in this country; early lung cancer detection has not resulted in improved survival rates. 1-5 These facts are disturbing when compared with the advances made .in the treatment Received August 3, 1990. Accepted May 24, 1991. From the Health Sciences Program, SRI International, Menlo Park, CA; Cytopathology Division, Department of Pathology, University of Alberta, Edmonton; Pathology Laboratory Los Gatos, CA; and Lake Oswego, OR. Address reprint requests to G. Berry Schumann, M.D., Cytopathology Division, Department of Pathology, University of Alberta, 5B4.02 WC Mackenzie Health Centre, Edmonton, Alberta, Canada T6G 2R7. 0 1991 WILEY-LISS, INC.
of both cervical and breast disease after capabilities for early cytologic detection were developed. There is preliminary evidence, however, that detection of precursor stages of lung cancer, such as dysplasia (atypical metaplasia) or benign metaplasia, combined with vitamin therapy, may alter the progression of lung disease. 6-9 Even though some cellular transformations in the tracheobronchial tree associated with cigarette smoking have been described, little information is available regarding nonsmokers for comparative analysis. An analysis of nonsmokers could provide a framework to better understand the progressive steps to lung disease and to allow for early detection and intervention, thus changing the natural course of this disease. Although sputum cytopathology has long been used to detect lung cancer in the early and more advanced clinical states, little work has been done to understand and identify progressive cellular and noncellular changes that may precede the preneoplastic transformation of the epithelium lining the tracheobronchial tree. If we are to gain insight into the pathogenesis of lung disease, it is important to study these responses in an objective and standardized manner. Drawing from both human and animal models described by Fullmer, l 3 Saccomanno, Naisell, l 9 Becci,20 Keenan, 21-24 and McDowell 25 as a basis, our laboratory has developed a standardized and reliable approach to the analysis of material from the lungs using a quantitative form of sputum cytology. 26,27 The following eight cytomorphologic components associated with bronchial stimulation were included in this system: alveolar macrophages, pigmentation of macrophages, neutrophils, mucus, mucous spirals, columnar cells, metaplastic cells, and dysplastic cells. The primary objective of this study was to ascertain the extent and nature of cytomorphologic differences in sputum collected from smokers and nonsmokers using these eight components.
''
'*
Diagnostic Cytopathology, Vol 7, No 6
569
SWAN ET AL.
Materials and Methods A cross-sectional analysis of cytologic data from adequate sputum specimens gathered from 349 cigarette smokers and 93 nonsmokers from January 1987 through March 1988 is presented. Sputum samples were obtained from smoking subjects who were patients of pulmonologists and primary care physicians; samples from subjects under treatment for serious respiratory ailments (e.g., cancer, COPD, pneumonia, asthma, and upper respiratory infections) were excluded from this analysis. Smoking histories revealed that these were moderate smokers (48.1 30.1 packyears). The nonsmoking group consisted ofvolunteers affiliated with the Church of Jesus Christ of Latter Day Saints located in the San Francisco Bay area. Both nonsmokers and smokers were middle-aged subjects (Table I). The subjects were provided with a specialized sputum collection container, specifically designed to collect and transport samples by mail. The containers were prefilled with Saccomanno’s fixative. A questionnaire attached to the sputum cytology request form required a brief health history on the patient, including respiratory symptoms (productive cough, chest pain, wheezing, and shortness of breath), occupational exposure to toxins, and smoking history. The patient was instructed to collect an earlymorning, 3-day pooled sputum specimen by deep coughing and expectorating into the container. Judged by the number of macrophages per slide, this technique has been shown to yield satisfactory samples for cytologic analysis in 90% of smokers and 85% of nonsmokers. Specimens were processed by the Saccomanno blended method: homogenization at 22,000 rpm for 5-20 sec and centrifugation at 1,500 rpm for 10 min. The supernatant was decanted and the remaining material reconstituted with an equal part of 50% ethyl alcohol. The specimen was resuspended and 75 p1were aliquoted onto a glass micro-
+
Table I. Selected Characteristics for Nonsmokers and Smokers
Age M (SD) Gender Male N (%) Female N (%) Occupational exposure Yes N (%) No N (9’0) Respiratory symptoms Yes N (%) NO N (96) Packyears M (SD)
Nonsmokers (N = 93)
Smokers (N = 349)
P value
44.9
(14.8)
51.5
(13.6)
An exploratory analysis of sputum cytology was conducted to determine the extent and nature of cytomorphologic diferences between smokers and nonsmokers. Specijically, alveolar macrophages, pigmen tation of macrophages, n eu trophils, m ucus, mucous spirals, columnar cells, metaplastic cells, and dysplastic cells were analyzed and compared in 349 smokers and 93 nonsmokers. Results indicate that smokers have signijicantly higher scores on all components than do nonsmokers. Pigmentation of macrophages, alveolar macrophages, and mucous spirals were the major components discriminating between smokers and nonsmokers. A comparison of the nonsmoking sample with a diferent group of smokers (n = 466) revealed similar findings. Using the cytomorphologic parameters as the basis for classifcation, discriminant function analysis was able to classify correctly 94 percent of the nonsmokers and 91 % of the smokers, a finding confirmed in the second group of smokers. These data indicate that smokers d@er reliably from nonsmokers with respect to their cytomorphologic profile. Diagn Cytopathol 1991;7:569-575. Key Words: Sputum cytology; Cigarette smoking; Quantitative
sputum cytology Cigarette smoke is the leading known cause of primary lung cancer and chronic obstructive pulmonary disease (COPD). The field of cytopathology has made significant contributions over the years to the detection of lung lesions in their earliest stages. However, the mortality rates for both lung cancer and COPD continue to rise in this country; early lung cancer detection has not resulted in improved survival rates. 1-5 These facts are disturbing when compared with the advances made .in the treatment Received August 3, 1990. Accepted May 24, 1991. From the Health Sciences Program, SRI International, Menlo Park, CA; Cytopathology Division, Department of Pathology, University of Alberta, Edmonton; Pathology Laboratory Los Gatos, CA; and Lake Oswego, OR. Address reprint requests to G. Berry Schumann, M.D., Cytopathology Division, Department of Pathology, University of Alberta, 5B4.02 WC Mackenzie Health Centre, Edmonton, Alberta, Canada T6G 2R7. 0 1991 WILEY-LISS, INC.
of both cervical and breast disease after capabilities for early cytologic detection were developed. There is preliminary evidence, however, that detection of precursor stages of lung cancer, such as dysplasia (atypical metaplasia) or benign metaplasia, combined with vitamin therapy, may alter the progression of lung disease. 6-9 Even though some cellular transformations in the tracheobronchial tree associated with cigarette smoking have been described, little information is available regarding nonsmokers for comparative analysis. An analysis of nonsmokers could provide a framework to better understand the progressive steps to lung disease and to allow for early detection and intervention, thus changing the natural course of this disease. Although sputum cytopathology has long been used to detect lung cancer in the early and more advanced clinical states, little work has been done to understand and identify progressive cellular and noncellular changes that may precede the preneoplastic transformation of the epithelium lining the tracheobronchial tree. If we are to gain insight into the pathogenesis of lung disease, it is important to study these responses in an objective and standardized manner. Drawing from both human and animal models described by Fullmer, l 3 Saccomanno, Naisell, l 9 Becci,20 Keenan, 21-24 and McDowell 25 as a basis, our laboratory has developed a standardized and reliable approach to the analysis of material from the lungs using a quantitative form of sputum cytology. 26,27 The following eight cytomorphologic components associated with bronchial stimulation were included in this system: alveolar macrophages, pigmentation of macrophages, neutrophils, mucus, mucous spirals, columnar cells, metaplastic cells, and dysplastic cells. The primary objective of this study was to ascertain the extent and nature of cytomorphologic differences in sputum collected from smokers and nonsmokers using these eight components.
''
'*
Diagnostic Cytopathology, Vol 7, No 6
569
SWAN ET AL.
Materials and Methods A cross-sectional analysis of cytologic data from adequate sputum specimens gathered from 349 cigarette smokers and 93 nonsmokers from January 1987 through March 1988 is presented. Sputum samples were obtained from smoking subjects who were patients of pulmonologists and primary care physicians; samples from subjects under treatment for serious respiratory ailments (e.g., cancer, COPD, pneumonia, asthma, and upper respiratory infections) were excluded from this analysis. Smoking histories revealed that these were moderate smokers (48.1 30.1 packyears). The nonsmoking group consisted ofvolunteers affiliated with the Church of Jesus Christ of Latter Day Saints located in the San Francisco Bay area. Both nonsmokers and smokers were middle-aged subjects (Table I). The subjects were provided with a specialized sputum collection container, specifically designed to collect and transport samples by mail. The containers were prefilled with Saccomanno’s fixative. A questionnaire attached to the sputum cytology request form required a brief health history on the patient, including respiratory symptoms (productive cough, chest pain, wheezing, and shortness of breath), occupational exposure to toxins, and smoking history. The patient was instructed to collect an earlymorning, 3-day pooled sputum specimen by deep coughing and expectorating into the container. Judged by the number of macrophages per slide, this technique has been shown to yield satisfactory samples for cytologic analysis in 90% of smokers and 85% of nonsmokers. Specimens were processed by the Saccomanno blended method: homogenization at 22,000 rpm for 5-20 sec and centrifugation at 1,500 rpm for 10 min. The supernatant was decanted and the remaining material reconstituted with an equal part of 50% ethyl alcohol. The specimen was resuspended and 75 p1were aliquoted onto a glass micro-
+
Table I. Selected Characteristics for Nonsmokers and Smokers
Age M (SD) Gender Male N (%) Female N (%) Occupational exposure Yes N (%) No N (9’0) Respiratory symptoms Yes N (%) NO N (96) Packyears M (SD)
Nonsmokers (N = 93)
Smokers (N = 349)
P value
44.9
(14.8)
51.5
(13.6)
