Malignant pleural effusions can be a consequence of primary lung malignancies (including mesothelioma) or metastases from extrathoracic malignancies. More than 75% of malignant pleural effusions are caused by malignant processes involving the lung, breast, ovary or lymphomas.1-4 Metastatic adenocarcinoma is the most common tumour type.5Malignant pleural effusion is an indicator of poor prognosis as the presence of a malignant pleural effusion upstages a patient to Stage IV disease in non small cell lung cancer or extensive disease in small cell lung cancer.The median survival from clinical recognition is 4 months irrespective of the cause of the malignant pleural effusion.6 However, prolonged survival is possible in some patients.Accurate and prompt diagnosis is important in determining the best management options for patients with malignant pleural effusions. Sending pleural fluid for cytological analysis is the usual first step. Previous studies have shown that the accuracy of pleural fluid cytology in diagnosing malignant pleural effusions varies from centre to centre and is reported to be between 40% and 87%7-13. The diagnostic yield increases with repeated pleural fluid aspirations but is not dependent on the volume of pleural fluid submitted for cytologic analysis14-15.The first objective of this audit was to evaluate the diagnostic yield of pleural fluid cytology for patients admitted to Middlemore Hospital (Auckland, New Zealand) with malignant pleural effusions over a 1-year periodThe second objective was to document the waiting time for cytology results to return to the admitting team and whether this contributed to length of stay.The third objective was to evaluate whether the volume of pleural fluid analysed contributed to diagnostic yield.Methods Patient population and study protocol All patients age \u226515 admitted under the Respiratory Services at Middlemore Hospital and diagnosed with malignant pleural effusions between 31 May 2010 and 1 June 2011 were included. Patients with pleural effusion from infection or fluid overload were excluded. Data was collected retrospectively by reviewing paper and electronic medical records. Patient demographics, investigations (including time pleural fluid sent to the lab, time results available, volume of pleural fluid analysed for cytology), clinical diagnosis, pathological diagnosis, admission and discharge details and death were extracted. A positive pleural fluid cytology was defined as a sample containing abnormal cells confirming malignancy. Results Patient characteristics Thirty-six patients (21 male, 15 female) were admitted to Middlemore Hospital with malignant pleural effusions over the 1 year period of study. The age range was 36-92, median age 67. Ethnicity of patients included European (23), Maori (6), Pacific (5) and Asian (2). Seventeen patients had a known diagnosis of cancer before admission with malignant pleural effusion. The most common cause of malignant pleural effusion was lung cancer (55.6 %), the most common of which were adenocarcinoma (25%), mesothelioma (19%), and small cell lung carcinoma (8.3%). Breast cancer (13.9%) and gynaecological cancers (11.4%) including 3 ovarian cancers were the next most common (Table 1). Table 1. Types of malignancy in patients with malignant pleural effusions Investigations Thirty one out of 36 patients had pleural fluid sent for analysis. Five werent sent as a prior diagnosis of malignancy was known and the diagnosis was presumptive based on the clinical and radiological features. 54.8% (17/31) had positive pleural fluid cytology, 16 of whom were diagnosed on the initial aspirate (51.6%). Eight of the 16 had a pre-existing diagnosis of malignancy. Only 4 of the 15 patients with negative initial pleural fluid cytology had a repeat aspiration (1 of which was positive). Thirteen out of 31 patients did not require further investigations (12 had a known diagnosis of malignancy before MPE and 1 did not want further investigations). Of the remaining 18 patients without a prior diagnosis of malignancy, 33% (6/18) of patients had diagnosis made with pleural fluid cytology alone. Other diagnostic techniques include 6 ultrasound guided biopsies (33%), 2 CT guided biopsies (11%), 2 medical thoracoscopies (11%), 2 bronchoscopes (11%) and 2 blind biopsies (11%). 2 patients had 2 different biopsies (1 had an ultrasound guided pleural biopsy which was non-diagnostic then medical thoracoscopy while another had a fine needle aspiration of supraclavicular mass and bronchoscopy with right lower lobe biopsies). Cytology turnaround times and length of hospital stay Overall, the average cytology turnaround time was 8.69 days with a range of 2.23-43.06 days (Figure 1). For the 16 samples with positive pleural fluid cytology, the average turnaround time was 11 days (range 3.49-43.06 days) while the 15 samples with negative cytology had an average turnaround time of 6.23 days (range 2.23-11.11 days). The day of the week pleural aspirates were obtained did not appear to influence the turnaround time. Figure 1. Plot of cytology turnaround times The length of hospital stay ranged from 1.11-20.8 days. Average length of stay was 7.77 days. Seventeen patients had cytology results available before discharge while 14 patients did not. Table 2 summarises the comparison between these two groups. In the former group, the average cytology turnaround time was shorter than the latter group (6.23 days compared to 11.68 days) but the average length of hospital stay was longer (10.96 days compared to 5.2 days). 35% (6/17) of the former group had a chemical pleurodesis as part of their admission compared only 7% (1/14) in the latter group. The majority of the former group had negative pleural fluid cytology and did not have a known diagnosis of malignancy before admission with MPE whereas the majority of the latter group had positive pleural fluid cytology. Table 2. Comparison of group discharged after cytology results available and group discharged before cytology results available Variables Discharge after cytology results available (n=17) Discharge before cytology results available (n=14) Positive cytology Negative cytology Known diagnosis before MPE Diagnosis not known Cytology waiting time: average (days) Cytology waiting time: range (days) Length of hospital stay: average (days) Length of hospital stay: range (days) 6 11 5 12 6.23 2.23-12.93 10.96 1.11-16.06 10 4 7 7 11.68 3.7-43.06 5.2 2.92-20.8 Table 3 summarises the comparison between patients with a known diagnosis of malignancy before admission with MPE and patients without. The latter group had a higher proportion of negative initial pleural fluid cytology and seemed to have a shorter average cytology turnaround time (7.95 days compared to 9.87 days) but a longer average length of hospital stay (9.55 days compared to 6.46 days). A higher proportion of the former group was discharged before cytology results. Table 3. Comparison of group with known diagnosis of cancer before MPE and group without Variables Diagnosis known (n=12) Diagnosis unknown (n=19) Positive cytology Negative cytology Discharge before cytology results Discharge after cytology results Cytology waiting time: average (days) Cytology waiting time: range (days) Length of hospital stay: average (days) Length of hospital stay: range (days) 8 4 7 5 9.87 2.23-43.06 6.46 1.11-16.06 8 11 7 12 7.95 3.49-16.14 9.55 1.99-20.8 Pleural fluid volume analysed for cytology The volume of pleural fluid analysed for cytology ranged between 2 to 2105mL. It was not documented in 5 samples (2 with positive cytology and 3 without). Seventeen samples were \u2264 50mL with a mean volume of 19mL (range 2 to 40mL). 41.2% (7/17) had positive cytology. Nine samples were >50mL with a mean volume of 452mL (range 54 to 2105mL). 77.8% (7/9) had positive cytology (Table 4). However, the difference in diagnostic yield between the two groups was not statistically significant (p=0.08). Table 4. Comparison of pleural fluid specimens \u2264 50mL to specimens > 50mL All (n=26*) Positive cytology Negative cytology \u2264 50mL pleural fluid analysed > 50mL pleural fluid analysed 7 7 10 2 *5 samples were excluded as they had no volumes documented (2 had positive cytology and 3 without). Survival following admission with malignant pleural effusion Overall, there were 31 deaths. 5 patients are still alive as of December 2012 (1 lung adenocarcinoma, 1 mesothelioma, 3 breast cancers). Survival ranged from 8 to 467 days. Median survival was 124 days, mean survival was 163.81 days. Discussion The diagnostic yield of initial pleural fluid cytology ranged from 48.5% to 63% on documented studies10,12,13 and improves with repeated samples. The diagnostic yield of initial pleural fluid cytology for our cohort of patients was 51.6%, improving to 54.8% with repeat sampling. This is in line with documented studies. There is a large variation in cytology turnaround times. Samples with positive cytology have longer turnaround times on average. This presumably was because positive results needed special staining and further processing/identification before they are made available. Interestingly, the cytology turnaround times were shorter for patients who remained in hospital compared to those discharged before the results were available. This could be due to the inpatient team actively contacting the lab to expedite results. However, a higher proportion of those who remained in hospital had negative initial pleural fluid cytology. The length of hospital stay does not seem to directly correlate with cytology turnaround times as the average cytology turnaround times were shorter for those who remained in hospital. Instead, longer hospital stays were seen more in patients with negative initial pleural fluid cytology and those with no prior diagnosis of cancer suggesting that these patients were staying in hospital for further investigations to establish a formal diagnosis which could influence management decisions. Other reasons that could potentially explain prolonged hospital stay (e.g. patients more unwell, complex discharge planning etc) were not examined in this audit. Sallach et al14 and Abouzgheib et al15 suggest that the diagnostic yield is not dependent on volume of pleural fluid submitted for cytology. Whilst evaluating whether pleural fluid volume influenced diagnostic yield was not a primary outcome and the number of patients studied is too small to test this hypothesis, our results favoured sending samples of >50ml although the difference in yield was not significant (p=0.08). Finally, the median survival of our cohort of patients was 124 days. This is consistent with the findings of 4 months from Heffner et al6. Conclusion Diagnostic yield from pleural fluid cytology at our hospital is comparable with other documented studies. ALOS appears to be influenced by a negative initial pleural fluid cytology and the uncertainty of diagnosis of cancer, not cytology turnaround time. Our findings suggest that a more efficient diagnostic and treatment algorithm could be considered with emphasis on Day Stay investigation and treatment as apart from waiting for results of pleural aspirate, there were few if any other management decisions taken and few other staging or diagnostic procedures other than when cytology was negative.

Malignant pleural effusions can be a consequence of primary lung malignancies (including mesothelioma) or metastases from extrathoracic malignancies. More than 75% of malignant pleural effusions are caused by malignant processes involving the lung, breast, ovary or lymphomas.1-4 Metastatic adenocarcinoma is the most common tumour type.5Malignant pleural effusion is an indicator of poor prognosis as the presence of a malignant pleural effusion upstages a patient to Stage IV disease in non small cell lung cancer or extensive disease in small cell lung cancer.The median survival from clinical recognition is 4 months irrespective of the cause of the malignant pleural effusion.6 However, prolonged survival is possible in some patients.Accurate and prompt diagnosis is important in determining the best management options for patients with malignant pleural effusions. Sending pleural fluid for cytological analysis is the usual first step. Previous studies have shown that the accuracy of pleural fluid cytology in diagnosing malignant pleural effusions varies from centre to centre and is reported to be between 40% and 87%7-13. The diagnostic yield increases with repeated pleural fluid aspirations but is not dependent on the volume of pleural fluid submitted for cytologic analysis14-15.The first objective of this audit was to evaluate the diagnostic yield of pleural fluid cytology for patients admitted to Middlemore Hospital (Auckland, New Zealand) with malignant pleural effusions over a 1-year periodThe second objective was to document the waiting time for cytology results to return to the admitting team and whether this contributed to length of stay.The third objective was to evaluate whether the volume of pleural fluid analysed contributed to diagnostic yield.Methods Patient population and study protocol All patients age \u226515 admitted under the Respiratory Services at Middlemore Hospital and diagnosed with malignant pleural effusions between 31 May 2010 and 1 June 2011 were included. Patients with pleural effusion from infection or fluid overload were excluded. Data was collected retrospectively by reviewing paper and electronic medical records. Patient demographics, investigations (including time pleural fluid sent to the lab, time results available, volume of pleural fluid analysed for cytology), clinical diagnosis, pathological diagnosis, admission and discharge details and death were extracted. A positive pleural fluid cytology was defined as a sample containing abnormal cells confirming malignancy. Results Patient characteristics Thirty-six patients (21 male, 15 female) were admitted to Middlemore Hospital with malignant pleural effusions over the 1 year period of study. The age range was 36-92, median age 67. Ethnicity of patients included European (23), Maori (6), Pacific (5) and Asian (2). Seventeen patients had a known diagnosis of cancer before admission with malignant pleural effusion. The most common cause of malignant pleural effusion was lung cancer (55.6 %), the most common of which were adenocarcinoma (25%), mesothelioma (19%), and small cell lung carcinoma (8.3%). Breast cancer (13.9%) and gynaecological cancers (11.4%) including 3 ovarian cancers were the next most common (Table 1). Table 1. Types of malignancy in patients with malignant pleural effusions Investigations Thirty one out of 36 patients had pleural fluid sent for analysis. Five werent sent as a prior diagnosis of malignancy was known and the diagnosis was presumptive based on the clinical and radiological features. 54.8% (17/31) had positive pleural fluid cytology, 16 of whom were diagnosed on the initial aspirate (51.6%). Eight of the 16 had a pre-existing diagnosis of malignancy. Only 4 of the 15 patients with negative initial pleural fluid cytology had a repeat aspiration (1 of which was positive). Thirteen out of 31 patients did not require further investigations (12 had a known diagnosis of malignancy before MPE and 1 did not want further investigations). Of the remaining 18 patients without a prior diagnosis of malignancy, 33% (6/18) of patients had diagnosis made with pleural fluid cytology alone. Other diagnostic techniques include 6 ultrasound guided biopsies (33%), 2 CT guided biopsies (11%), 2 medical thoracoscopies (11%), 2 bronchoscopes (11%) and 2 blind biopsies (11%). 2 patients had 2 different biopsies (1 had an ultrasound guided pleural biopsy which was non-diagnostic then medical thoracoscopy while another had a fine needle aspiration of supraclavicular mass and bronchoscopy with right lower lobe biopsies). Cytology turnaround times and length of hospital stay Overall, the average cytology turnaround time was 8.69 days with a range of 2.23-43.06 days (Figure 1). For the 16 samples with positive pleural fluid cytology, the average turnaround time was 11 days (range 3.49-43.06 days) while the 15 samples with negative cytology had an average turnaround time of 6.23 days (range 2.23-11.11 days). The day of the week pleural aspirates were obtained did not appear to influence the turnaround time. Figure 1. Plot of cytology turnaround times The length of hospital stay ranged from 1.11-20.8 days. Average length of stay was 7.77 days. Seventeen patients had cytology results available before discharge while 14 patients did not. Table 2 summarises the comparison between these two groups. In the former group, the average cytology turnaround time was shorter than the latter group (6.23 days compared to 11.68 days) but the average length of hospital stay was longer (10.96 days compared to 5.2 days). 35% (6/17) of the former group had a chemical pleurodesis as part of their admission compared only 7% (1/14) in the latter group. The majority of the former group had negative pleural fluid cytology and did not have a known diagnosis of malignancy before admission with MPE whereas the majority of the latter group had positive pleural fluid cytology. Table 2. Comparison of group discharged after cytology results available and group discharged before cytology results available Variables Discharge after cytology results available (n=17) Discharge before cytology results available (n=14) Positive cytology Negative cytology Known diagnosis before MPE Diagnosis not known Cytology waiting time: average (days) Cytology waiting time: range (days) Length of hospital stay: average (days) Length of hospital stay: range (days) 6 11 5 12 6.23 2.23-12.93 10.96 1.11-16.06 10 4 7 7 11.68 3.7-43.06 5.2 2.92-20.8 Table 3 summarises the comparison between patients with a known diagnosis of malignancy before admission with MPE and patients without. The latter group had a higher proportion of negative initial pleural fluid cytology and seemed to have a shorter average cytology turnaround time (7.95 days compared to 9.87 days) but a longer average length of hospital stay (9.55 days compared to 6.46 days). A higher proportion of the former group was discharged before cytology results. Table 3. Comparison of group with known diagnosis of cancer before MPE and group without Variables Diagnosis known (n=12) Diagnosis unknown (n=19) Positive cytology Negative cytology Discharge before cytology results Discharge after cytology results Cytology waiting time: average (days) Cytology waiting time: range (days) Length of hospital stay: average (days) Length of hospital stay: range (days) 8 4 7 5 9.87 2.23-43.06 6.46 1.11-16.06 8 11 7 12 7.95 3.49-16.14 9.55 1.99-20.8 Pleural fluid volume analysed for cytology The volume of pleural fluid analysed for cytology ranged between 2 to 2105mL. It was not documented in 5 samples (2 with positive cytology and 3 without). Seventeen samples were \u2264 50mL with a mean volume of 19mL (range 2 to 40mL). 41.2% (7/17) had positive cytology. Nine samples were >50mL with a mean volume of 452mL (range 54 to 2105mL). 77.8% (7/9) had positive cytology (Table 4). However, the difference in diagnostic yield between the two groups was not statistically significant (p=0.08). Table 4. Comparison of pleural fluid specimens \u2264 50mL to specimens > 50mL All (n=26*) Positive cytology Negative cytology \u2264 50mL pleural fluid analysed > 50mL pleural fluid analysed 7 7 10 2 *5 samples were excluded as they had no volumes documented (2 had positive cytology and 3 without). Survival following admission with malignant pleural effusion Overall, there were 31 deaths. 5 patients are still alive as of December 2012 (1 lung adenocarcinoma, 1 mesothelioma, 3 breast cancers). Survival ranged from 8 to 467 days. Median survival was 124 days, mean survival was 163.81 days. Discussion The diagnostic yield of initial pleural fluid cytology ranged from 48.5% to 63% on documented studies10,12,13 and improves with repeated samples. The diagnostic yield of initial pleural fluid cytology for our cohort of patients was 51.6%, improving to 54.8% with repeat sampling. This is in line with documented studies. There is a large variation in cytology turnaround times. Samples with positive cytology have longer turnaround times on average. This presumably was because positive results needed special staining and further processing/identification before they are made available. Interestingly, the cytology turnaround times were shorter for patients who remained in hospital compared to those discharged before the results were available. This could be due to the inpatient team actively contacting the lab to expedite results. However, a higher proportion of those who remained in hospital had negative initial pleural fluid cytology. The length of hospital stay does not seem to directly correlate with cytology turnaround times as the average cytology turnaround times were shorter for those who remained in hospital. Instead, longer hospital stays were seen more in patients with negative initial pleural fluid cytology and those with no prior diagnosis of cancer suggesting that these patients were staying in hospital for further investigations to establish a formal diagnosis which could influence management decisions. Other reasons that could potentially explain prolonged hospital stay (e.g. patients more unwell, complex discharge planning etc) were not examined in this audit. Sallach et al14 and Abouzgheib et al15 suggest that the diagnostic yield is not dependent on volume of pleural fluid submitted for cytology. Whilst evaluating whether pleural fluid volume influenced diagnostic yield was not a primary outcome and the number of patients studied is too small to test this hypothesis, our results favoured sending samples of >50ml although the difference in yield was not significant (p=0.08). Finally, the median survival of our cohort of patients was 124 days. This is consistent with the findings of 4 months from Heffner et al6. Conclusion Diagnostic yield from pleural fluid cytology at our hospital is comparable with other documented studies. ALOS appears to be influenced by a negative initial pleural fluid cytology and the uncertainty of diagnosis of cancer, not cytology turnaround time. Our findings suggest that a more efficient diagnostic and treatment algorithm could be considered with emphasis on Day Stay investigation and treatment as apart from waiting for results of pleural aspirate, there were few if any other management decisions taken and few other staging or diagnostic procedures other than when cytology was negative.