Table 1LUNG CANCER: signs and symptoms, including risk factors

AuthorSettingDescriptionNo.InclusionExclusionResultsGold StdQuality
Alberg et al, 2003The study provided a summary of the epidemiological evidence on lung cancer.A single etiologic agent, cigarette smoking, has been noted as by far the leading cause of lung cancer accounting for approximately 90% of cases in the United States. Risk of lung cancer among cigarette smokers increases with the duration of smoking and the number of cigarettes smoked per day. This observation has been made repeatedly in cohort and case-control studies. Asbestos exposure may pose a risk to building occupants and radon has been associated with lung cancer.

The likelihood of developing lung cancer decreases among those who quit smoking compared to those who continue to smoke. As the period of abstinence from smoking cigarettes increases, the risk of lung cancer decreases. However, even for periods of abstinence of >40 years, the risk of lung cancer among former smokers remains elevated compared to never smokers. Studies have shown comparable reductions in risk following smoking cessation, regardless of sex, type of tobacco smoked and histologic type of lung cancer.
Almost one quarter of lung cancer cases among never-smokers are estimated to be attributed to exposure to passive smoking. Estimates derived from case-control studies of the proportion of lung cancer that is contributed to by occupational exposures have ranged widely, but most point estimates or ranges have included values from 9 to 15%.
A comprehensive review and systematic synthesis was not undertaken.
BTS 2001Guidelines on management of malignany mesotheliomaNot evidence based
DoH 2000Guidelines on referral for suspected cancerNationally recognised guidelines
G.I.V.I.O 1989ItalyRetrospective case series. Signs and symptoms were reported in a study of diagnostic and therapeutic care.380 patients from 20 hospitalsPatients with lung cancer seen in Italian general hospitals between January June 1987 irrespective of their age, sex and severity of diseaseSymptoms most frequently reported at presentation were cough in 175 (46%), shortness of breath in 86 (23%), chest pain in 87 (23%), haemoptysis in 75 (20%) and fever in 52 (14%).
Finally, 26 (9%) patients had symptoms due to distant metastases at diagnosis.
Histo/cytologic findings were available for 363 cases.The study did not distinguish between late or early symptoms.
The six month period of data collection was very short.
Herth et al 2001UKRetrospective case series. Cases of lung cancer were reported in patients presenting with haemoptysis of unknown origin722In 135 patients (19%) no aetiology for the bleeding could be determined and this group was targeted for further follow-up. Follow-up data were available for 20 patients. Eighty-one patients (60%) were smokers, 16 patients (12%) had a history of chronic obstructive pulmonary disease (COPD) and 10 patients (7%) had a history of tuberculosis.
Lung cancer developed in 7 of the 115 patients with unknown etiology despite unrevealing bronchoscopy and normal chest radiographic findings at initial presentation.
Using the cohort study analysis for unpaired differences, a 10% probability was found for lung cancer developing after haemoptysis of unknown origin if the patient was a current smoker and > 40 years old.
CytologyA sizeable patient population was followed for a reasonable length of time.
Koyi et al 2002SwedenProspective cohort study examined patients referred to a specialised centre. GPs were encouraged to refer all suspected cases of lung carcinoma including those with a seemingly dismal prognosis as early as possible. Definite diagnosis was aimed at with a biopsy and/or cytology test/-and in cases where this was not possible diagnosis was based on x-ray findings, clinical data and symptoms362All patients referred to a specialised centre between January 1997 and December 1999In 50 of the 364 patients (13.7%) biopsy and/or a cytology test was not possible due to patients’ unwillingness or ethical reasons. X-ray findings clinical data & symptoms were used instead.Data was provided on the initial symptoms of lung cancer prompting concern among patients and those which led them to consult a doctor. Good prospective study.
Krech et al 1992USAProspective cohort study.
Symptoms experienced by patients with advanced cancer were described using a standard assessment tool for one and a half years since October 1987
100Most common and severe symptoms were pain (86), dyspnoea (70) and anorexia (68). Males aged 64 experienced more easy fatigue (p=0.01), taste changes (p=0.009) and sleep problems (p=0.004), higher incidences of cough and > 10% weight loss. Nausea was more frequent in females (p=0.07) and wheezing in males (p=0.06), although neither was a dominant symptom.No reference to gold standard testThe data indicated advanced symptoms.
Different interpretations of weakness and fatigue may have affected the results reported.
Liedekerken et al 1997NetherlandsSystematic review examining the relationship between prolonged coughing and the diagnosis of lung cancer.Studies examining the relationship between prolonged coughing and lung cancer diagnosis.Studies were excluded if there were insufficient data for the calculations to be made or if patients were chosen selectively, other than by setting.No primary care could be identified. One paper reported the relationship between prolonged cough and lung cancer based on 6027 patients in a specialised setting. It revealed a high negative (0.99) and a low positive (0.03) predictive value, a sensitivity of 0.48 and a specificity of 0.71.--Thorough attempt was made to identify evidence on the significance of prolonged cough for lung cancer but scarcely any studies came to light.
Macbeth et al, 1996The risk factors associated with lung cancer have been identified as including tobacco, asbestos and radon. The influence of genetic factors and the effects of chromosomal abnormalities has also been assessed. At least thirty retrospective and eight prospective studies have established a link between cigarette smoking and lung cancer. It has been estimated that 85–90% of all lung cancers can be linked to active smoking.
The use of cigarettes carries a significantly greater risk of developing lung cancer than either pipe or cigar smoking.
The age of starting cigarette smoking, the duration of smoking and the nicotine content of the cigarettes are all important factors. The risk of lung cancer at the age of 60 years is reported to be three times greater for those who started smoking between the ages of 14 and 16 years compared to those who began 10 years later. It has been calculated that someone aged 35 years who smokes 25 or more cigarettes per day has a 13% chance of dying from lung cancer before the age of 75 years. Exposure to known carcinogens including asbestos, radon, chromium, nickel and inorganic arsenic compounds increases the risk of lung cancer.
Even a short exposure may be sufficient to cause lung cancer, if the concentration of asbestos is high enough. Miners who are exposed to high concentrations of radon have an increased risk of lung cancer, but its role in domestic housing as a factor causing lung cancer is uncertain.
Several studies have shown an increased risk in the siblings of patients who develop lung cancer.
Mansson et al 1994SwedenRetrospective case series. The records of patients with lung cancer reported to the Swedish Cancer Registry 1980–1984 were examined using hospital records, with special reference to GPs’ role40Records of all subjects with lung cancer reported to the Swedish Cancer Registry 1980–1984The mean and median ages at the time of the diagnosis was 69 and the range was 43–85 years.
The initial symptoms were cough followed by dyspnoea, chest pain, fever, weight loss and tiredness. Other presenting symptoms were oedema, haemoptysis, facial pain, pricking sensations in the throat, stuffed nose, dizziness, frequent colds and tumour outside the throat.
Symptoms included palpable lymph nodes (2 patients), dyspnoea, liver enlargement, cachexia, tendency to fall and an episode of unconsciousness. No abnormal signs were found on physical examination in 10 patients (26%).

The most common abnormal laboratory finding was increased erythrocyte sedimentation rate (>30mm/h) found in 17 of the 35 patients (49%) in whom it was recorded. Increased leucocyte particle concentration was observed in nine of 31 patients (29%). Other abnormal laboratory findings were anaemia (two patients) and increased serum concentration of alkaline phosphatases (one patient). The mortality during the study period for this series was 97% (38 of 39 patients).
Diagnosis was confirmed by means of bronchoscopy, mediastinoscopy with cytology or at autopsyRelationship of signs and symptoms to pulmonary cancer was not statistically evaluated.
Mansson et al 2001SwedenRetrospective case series. Diagnostic activities were scrutinised and coded when malignancy was suspected. Information on diagnostic activities drawn from patient records was coded where a malignancy may have been a differential diagnosis for colorectal, breast, lung and prostate cancer.6812 patientsPulmonary diagnostic codes comprised the greatest part of the study (9422 codes corresponding to 65%). Most of these codes were assumed to be accounted for by infectious diseases in the upper airways.
C-reactive protein tests were taken 865 times and nasopharyngeal cultures 580 times. Blood haemoglobin and ESR were tested 822 and 579 times respectively. X-ray was performed 643 times. The yield of malignancy following chest X-ray was low, 0.4%.
No gold standard procedure stated to confirm diagnosis.The retrospective design may not have identified all the signs and symptoms that patients presented with.
Authors did not state the procedure used to confirm the diagnosis. No statistical tests were reported.
Melling et al 2002UKRetrospective case series. An analysis of the proportion of patients referred according to lung cancer guidelines was conducted in order to assess how different pathways resulted in varying management.362 patientsPatients randomly selected from a Yorkshire cancer registryAny patients that had missing case notes or were receiving private treatment or extra-regional care.47.8% of lung cancer patients presented to hospital with a chest x-ray diagnosis of lung cancer. A total of 148 patients in the ‘without chest x-ray diagnosis group’ were referred to hospital because of their symptoms but with no prior chest x-ray. 11.3% presented as self referrals to A&E and the remainder were referred without a diagnosis of lung cancer by other routes mainly via GPs.
80% of the ‘with diagnosis group’ presented to their GP with mainly lung related symptoms (cough, chest pain or infection, haemoptysis or dyspnoea) compared to 69 (46.6%, CI: 38.4%, 55.0%) of those without a diagnosis. Patients who did not present initially with a lung cancer diagnosis were less likely to receive specialist care (62%: 96%) or have histological confirmation (57.1%: 80.3%) or receive surgery or radical radiotherapy (6.9%: 13.9%).
Surgery, chemotherapy and palliative radiotherapy were all used most frequently in the ‘with chest x-ray diagnosis group’, but the difference was only significant for surgery (P=0.035).
57.1% of patients presenting without a chest x-ray had histological confirmation of malignancy compared to 80.3% who did.Study concluded that patients presenting to hospital without a suspicious CXR were less likely to have specialist care, histological confirmation and had lower rates of active treatment.
NICE 2004Guidelines on the diagnosis and treatment of lung cancerStill out for consultation
Ruano-Ravina et al, 2003A systematic literature review on risk factors was undertaken following a MEDLINE and EMBASE search from 1985 onwardsEditorials, commentaries and published articles less than 50 casesConcluded that risk of developing smoking-related lung cancer depended on several factors including duration of habit (number of cigarettes per day), age at initiation and type of tobacco. Passive smoking was considered a risk factor for lung cancer (RR reported to be approximately 1.5) although exposure was very difficult to measure. Many occupational groups identified as at risk. Individuals in contact with dust or microscopic particles (asbestos, wood dust, silica) at higher risk of developing lung cancer despite the effects of environmental pollution being complicated to assess. Survival rated as being better in women than men with incidence reported as being at around 65 years of age. Slight ethnic differences observed with higher mortality rates among African-Americans. Certain diseases raised risk of developing lung cancer such as tuberculosis, chronic obstructive pulmonary disease and silicosis. Family history of lung cancer was associated with a rise in risk.
One study of women showed that subjects reporting a family history of lung cancer had a 1.9 fold risk (95% CI 0.7–5.6) of developing lung cancer and those reporting a family history of cancer had a 1.8 fold risk of developing lung cancer (95% CI 1.0–3.2). Lung cancer was more common in families with record of breast & ovarian cancer.
Ecological studies lacked information on certain confounders such as tobacco use.
No study results were combined. Details were not provided as to how the quality of the studies was assessed.
Sarlani et al 2003USARetrospective case report and series identified from the literature.
The aim was to evaluate facial pain as a presenting symptom of non-metastatic lung cancer.
32The mean age at presentation was 54 years (range 34 to 78).
The vast majority of the patients were smokers or former smokers. The facial pain preceded the diagnosis of lung cancer by a mean of 9 months (range 1–48). Facial pain related to non-metastatic lung cancer was almost invariably unilateral, always ipsilateral to the tumour.
Eighteen of the 32 cases (56.25%) involved right sided pain and 12 (37.5%) left-sided pain. The pain most commonly affected the ear, the jaws and the temporal region. Pain in or around the ear was present in 20 of the 32 cases (62.5%) and jaw pain in 14 cases (43.75%).
Such pain was commonly misdiagnosed as atypical facial pain, dental pain or pain associated with temporamandibular disorders (TMD) or trigeminal neuralgia
Not statedMethodological details as to how patients were selected from the literature and how comprehensive or systematic the search was for relevant cases was not made explicit.
SIGN 2004Guidelines on lung cancerEvidence based.
Nationally recognised.
SIGN 2002Referral guidelines for suspected cancerNationally recognised.
Based on unpublished audits and other published literature.
Smith et al, 1995Meta-analysis to evaluate the relation between exposure to crystalline silica and lung cancer.29 studiesStudies for which effect measures (such as RRs and ORs) could be extracted for lung cancer mortality among SilicoticsStudies were excluded they were deemed to under or overestimate lung cancer riskAfter adjustment for competing risks, all 29 studies demonstrated lung cancer relative risk (RR) estimates greater than one. The pooled RR estimate for all studies that could be combined was 2.2, with a 95% CI of 2.1–2.4. The pooled estimates by study design were 2.0 (95% CI=1.8–3.3) for case-control studies. The proportional mortality studies combined gave a summary RR of 2.0 (95% CI=1.7–2.4) whereas the studies of cancer incidence gave a summary RR of 2.7 (95% CI = 2.3–3.2).It was not clear whether the search for relevant studies was systematic.
The potential confounding factor that could have exerted the most influence on results was smoking.
The issues surrounding the process of quality assessment of studies reviewed was not highlighted in detail.
Sridhar et al 1990USARetrospective case series. Hospital charts of patients with adenosquamous lung carcinoma identified between 1975 and 1988 were reviewed to determine the clinical features of lung cancer at the time of presentation.127Patients presented with cough 68 (54%); weight loss 54 (43%); expectoration 49 (39%); anorexia 45 (35%); chest pain 41 (32%); dyspnea 38 (30%); weakness 38 (30%); haemoptysis 30 (24%); pneumonia 16 (13%); fever 16 (13%); nausea 13 (10%); vomiting 9 (7%); dizziness 8 (6%); chills 6 (5).
All the study patients had histopathologic or cytologic diagnosis of adenosquamous carcinoma of the lung established by a Pathology Department. It was reported that haemoptysis was a more common presenting symptom in men than in women (p=0.05).
Histopathologic or cytologic diagnosisNo mention was made of follow up The value of digital clubbing in reinforcing suspicion of malignancy in those with non small cell lung cancer is highlighted.
Sridhar et al 1998USAProspective cohort study. The aim was to determine the relative frequency of clubbing in small cell lung carcinoma (SCLC) versus non-small cell lung carcinoma (NSCLC).
patients with a pathological diagnosis of lung cancer were examined for the presence or absence of digital clubbing.
Comparisons were made between patients with and without clubbing on the following: age, sex, substance use, tobacco, smoking history, family history of lung cancer and subtype of cancer
111 patientsClubbing was present in 32 (29%) of the 111 patients with lung cancer. Clubbing was more common in women (40%) than in men (19%; χ2 test p=0.011) and was more common in patients with NSCLC (35%) than those with SCLC (4%; χ2 test p=0.0036).PathologyDifficult to evaluate the frequency and occurrence of symptoms retrospectively.
Patients may not report all features unless asked
Tyczynski et al, 2003An epidemiological review of lung cancer in Europe reported risk related factorsTobacco smoking featured as the most prominent risk in developing lung cancer. A clear dose-response relation was reported to exist between lung-cancer risk and the number of cigarettes smoked per day, degree of inhalation and age at initiation of smoking. A person who has smoked all their life has a lung cancer risk 20–30 times greater than a non-smoker. Lung cancer risk decreases with time since smoking cessation.
The observation that the risk of lung cancer is greater in women than in men exposed to equivalent amounts of tobacco smoke is not supported by recent studies which conclude that risk is similar between the two sexes. Passive exposure to tobacco smoke also increases the risk of lung cancer. It is estimated that environmental exposure to tobacco smoke increases risk by 15–25%.
Additional factors contributing lung cancer risk includes increasing duration of exposure to asbestos which rises almost two-fold in those subjected to it the longest. A synergistic (multiplicative) effect between asbestos and tobacco smoking and the relation between these two factors was documented in three comprehensive reviews.
Occupational exposure to carcinogens and residential exposure to radon may increase the risk of lung cancer in men who never smoked. The combined effect of smoking and radon exposure however, is unknown.

From: Appendix A, Evidence Tables

Cover of Referral Guidelines for Suspected Cancer in Adults and Children
Referral Guidelines for Suspected Cancer in Adults and Children [Internet].
NICE Clinical Guidelines, No. 27.
Clinical Governance Research and Development Unit (CGRDU), Department of Health Sciences, University of Leicester.
Copyright © 2005, National Collaborating Centre for Primary Care.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.