Platinum versus non‐platinum chemotherapy regimens for small cell lung cancer
Abstract
Background
Small cell lung cancer (SCLC) is a very fast growing form of cancer and is characterised by early metastasis. As a result, chemotherapy is the mainstay of treatment. A number of different platinum‐based chemotherapy regimens and non‐platinum‐based chemotherapy regimens have been used for the treatment of SCLC, with varying results. This review was conducted to analyse the data from these studies in order to compare their effectiveness.
Objectives
To determine the effectiveness of platinum chemotherapy regimens compared with non‐platinum chemotherapy regimens in the treatment of SCLC with respect to survival, tumour response, toxicity and quality of life.
Search methods
We searched the biomedical literature databases CENTRAL (TheCochrane Library 2014, Issue 7), MEDLINE, EMBASE and CINAHL from 1966 to August 2014. In addition, we handsearched reference lists from relevant resources.
Selection criteria
All randomised controlled trials involving patients with pathologically confirmed SCLC (including both limited‐stage disease and extensive‐stage disease) and the use of a platinum‐based chemotherapy regimen in at least one treatment arm and a non‐platinum‐based chemotherapy regimen in a separate arm.
Data collection and analysis
We used standard methodological procedures expected by the Cochrane Collaboration. Two authors independently assessed search results. We assessed included studies for methodological quality and recorded the following outcome data: survival, tumour response, toxicity and quality of life. We combined the results of the survival, tumour response and toxicity data in a meta‐analysis. Quality‐of‐life data were analysed individually.
Main results
A total of 32 studies involving 6075 patients with SCLC were included in this systematic review. The majority of studies were multi‐centre randomised controlled trials conducted throughout Europe, North America and Asia with the earliest study publishing data in 1981 and the latest in 2014. The duration of studies ranged from 12 to 72 months with a median of 32 months. The median age of patients in the vast majority of studies was between 60 and 65 years of age. Eighteen studies presented data on extensive‐stage disease. Nine studies presented data on limited‐stage disease. Eleven studies did not present data based on the disease stage. These data were analysed separately in subgroup analyses. Sixteen (50%) studies were of good quality with a low risk of bias and the data from these studies were analysed separately in a heterogeneity analysis.
There was no statistically significant difference between treatment groups in terms of survival at 6 months, 12 months and 24 months. There was also no statistically significant difference in terms of overall tumour response. However, platinum‐based treatment regimens did have a significantly higher rate of complete response. Platinum‐based chemotherapy regimens had significantly higher rates of nausea and vomiting and thrombocytopenia toxicity. Four trials presented quality‐of‐life data, but, due to the different systems used to measure quality of life this data could not be combined in a meta‐analysis.
Authors' conclusions
Platinum‐based chemotherapy regimens did not offer a statistically significant benefit in survival or overall tumour response compared with non‐platinum‐based regimens. However, platinum‐based chemotherapy regimens did increase complete response rates, at the cost of higher adverse events including nausea and vomiting, anaemia and thrombocytopenia toxicity. These data suggest non‐platinum chemotherapy regimens have a more advantageous risk‐benefit profile. This systematic review highlights the lack of quality‐of‐life data in trials involving chemotherapy treatment for SCLC. With poor long‐term survival associated with both treatment groups, the issue of the quality of the survival period takes on even more significance. It would be beneficial for future trials in this area to include a quality‐of‐life assessment.
PICOs
Plain language summary
A comparison of platinum‐based and non‐platinum‐based chemotherapy regimens for the treatment of small cell lung cancer
Review question
Do patients with small cell lung cancer (SCLC) who receive platinum‐based chemotherapy treatment live longer than those who receive non‐platinum‐based chemotherapy treatment?
Other questions include: do these patients also respond better to treatment, experience fewer side‐effects and have a better quality of life?
Background
SCLC is a type of cancer that originates in the lungs. It is a very aggressive form of cancer that tends to grow and spread throughout the body quickly. As a result, chemotherapy is often the first type of treatment used for this type of cancer. Another type of treatment used for SCLC is radiotherapy, which is often given to the lung or to the brain.
A combination of a number of chemotherapy drugs used together is called a ‘chemotherapy regimen’. Currently, there are two main chemotherapy regimens used for treating SCLC:
• platinum‐based chemotherapy regimens – containing a chemotherapy drug known as a “platinum agent” in combination with other chemotherapy drugs and;
• non‐platinum‐based chemotherapy regimens – containing other chemotherapy drugs without a “platinum agent”.
Over the past years, many studies have been done comparing the use of platinum‐based chemotherapy regimens and non‐platinum‐based chemotherapy regimens in SCLC.
We carried out a study, called a meta‐analysis, which included patients with SCLC who took part in randomised controlled trials comparing platinum‐based chemotherapy regimens and non‐platinum‐based chemotherapy regimens.
Study characteristics
We searched for studies up to 1st August 2014. A total of 32 studies were part of this review and included 6,075 patients in total. The studies were carried out in many different countries throughout Europe, Asia and North America. The studies were conducted between 1981 and 2014.
Key results
The review showed that patients who received platinum‐based chemotherapy were not any more likely to be alive at 6 months, 12 months and 24 months after treatment compared with patients who received non‐platinum‐based chemotherapy.
Platinum‐based chemotherapy, however, showed higher rates of complete tumour response (the complete disappearance of tumours, at least for a period of time after treatment) compared to non‐platinum‐based chemotherapy. Platinum‐based chemotherapy also caused some more side effects, including nausea and vomiting, and low platelets.
Only four studies looked at quality of life but because they each used different methods to measure the effects, their results could not be combined. However, in each study there was no difference in the quality of life between the platinum‐based chemotherapy group and the non‐platinum‐based chemotherapy group.
Authors' conclusions
Background
Lung cancer is one of the most common cancers in the world, both in terms of incidence and mortality (Dela Cruz 2011). Smoking is the biggest risk factor for lung cancer and is associated with 90% of cases (Alberg 2003). Other risk factors include occupational and environmental exposure, for example asbestos (Heintz 2010). Lung cancer is divided into small cell lung cancer (SCLC) and non‐small cell lung cancer (NSCLC) based on histological appearance. SCLC makes up about 20% of lung cancers (Le Pechoux 2004).
SCLC is divided into two stages ‐ limited disease stage and extensive disease stage. Based on the Veterans' Administration Lung Study Group (VALG) classification, The International Association for the Study of Lung Cancer (IASLC) defined limited disease (LD‐SCLC) as: "Disease restricted to one hemithorax with regional lymph node metastases, including hilar, ipsilateral and contralateral mediastinal, and ipsilateral and contralateral supraclavicular nodes and should also include patients with ipsilateral pleural effusion independent of whether cytology is positive or negative" (Micke 2002).
Extensive disease (ED‐SCLC) was given the definition: "All patients with sites of disease beyond the definition of limited disease" (Micke 2002).
In SCLC, long‐term survival is quite poor. Untreated, the median survival is 4 to 12 weeks (Chan 2013). Because SCLC is a very aggressive type of cancer and early metastasis (both local and distant) is common, chemotherapy is the first line treatment (Stinchcombe 2010). Even with treatment, long‐term survival is poor. Median survival is in the order of 15 to 20 months for LD‐SCLC and 8 to 13 months for ED‐SCLC. Two‐ and five‐year survival rates for LD‐SCLC range from 20% to 40% and 10% to 13% respectively. For ED‐SCLC, survival rates are even poorer ‐ less than 5% at two years and 1% to 2% at five years (Glisson 2014).
Chemotherapy is the most common treatment for SCLC because of early metastatic spread. Platinum therapy has been widely used and is regarded as first line treatment, as it has been considered one of the most efficacious agents. It is often combined with the non‐platinum agent etoposide (Stinchcombe 2010). Platinum agents are cytotoxic alkylating agents that are active throughout the cell cycle (Chabner 2010). The most widely used platinum agents in SCLC are cisplatin (cis‐diamine‐dichloroplatinum II) and carboplatin (cis‐diamine‐(1,1‐cyclobutanedicarboxylate) platinum) (Stinchcombe 2010).
Non‐platinum agents for SCLC include vincristine, doxorubicin (Adriamycin), cyclophosphamide and ifosfamide. All of these agents have been shown to have anti‐tumour activity and have also been used in combination regimens in SCLC (Pujol 2000). One of the most common non‐platinum combinations that has been shown to be effective in SCLC is the vincristine, doxorubicin and cyclophosphamide regimen (Stinchcombe 2010).
As the available treatments have varying success rates and the different treatments have various advantages and disadvantages, a systematic review will be useful in determining optimal treatment regimens for SCLC.
Objectives
To determine the effectiveness of platinum chemotherapy regimens compared with non‐platinum chemotherapy regimens in the treatment of small cell lung cancer (SCLC) with respect to survival, tumour response, toxicity and quality of life.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs) that compared platinum‐based chemotherapy regimens with other non‐platinum‐based chemotherapy regimens.
Inclusion criteria:
-
Studies involving only patients with pathologically confirmed (cytological or histological) small cell lung cancer.
-
Studies with at least one platinum‐based treatment group.
-
Studies with at least one non‐platinum‐based treatment group.
Exclusion criteria:
-
Studies with platinum agents used in all treatment groups.
-
Studies with no platinum agent used in any treatment groups.
Types of participants
Adult patients of either sex with pathologically confirmed (cytological or histological) small cell lung cancer.
Types of interventions
-
Platinum agents at any dose or for any number of cycles compared with any other chemotherapy regimen [P versus A].
-
Platinum agents at any dose or for any number of cycles in combination with other chemotherapy regimens versus the same chemotherapy regimen without the platinum agent (i.e. non‐platinum chemotherapy is identical in both interventions) [(P+A) versus A].
-
Platinum agents at any dose or for any number of cycles in combination with any other chemotherapy regimen versus any other chemotherapy regimen not containing platinum agents [(P+A) versus B].
(Where P = platinum chemotherapy agents, A = non‐platinum chemotherapy regimens and B = non‐platinum chemotherapy regimens (different from A)).
Studies where platinum agents were administered to the control group were excluded from this review.
Radiotherapy
RCTs that involve the use of radiotherapy (RT) were included, provided that RT was planned to be given in an identical way (dose, fractionation, timing and technique) in both treatment arms. If RT was given unequally to a treatment arm, or if the chemotherapy regimen to which patients were randomised routinely affected the way in which the RT was given, then the RCT was excluded.
Types of outcome measures
The primary outcome measure was survival at 6 months, 12 months and 24 months of follow up. Other outcome measures, such as tumour response, treatment‐related toxicity and quality of life were also considered. Tumour response for objective overall response and complete response were defined as per World Health Organization (WHO) guidelines for tumour response evaluation (Park 2003). Toxic events were classified (if they had not already been) according to the WHO scale (WHO 1979) and only grades 3 and 4 of toxicity were analysed. The following toxic events were considered: toxic death, nausea and vomiting, alopecia, infection, anaemia, leukopenia, thrombocytopenia and granulocytopenia. Quality‐of‐life data were recorded qualitatively.
Search methods for identification of studies
Please see Appendix 1 and Appendix 2 for the search strategies used in the original review and current update, respectively
In 2007, we designed a search strategy to search the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2007, Issue 2), MEDLINE (accessed through PubMed), EMBASE and CINAHL (accessed through EBSCO). The date of the search was from 1966 to April 2007. The search strategy was not restricted by date or language.
The search strings used to retrieve studies are reported in Appendix 1
For this update, a search strategy generated by the Cochrane Lung Cancer Group was used to search the following databases:
-
Cochrane Central Register of Controlled Trials (CENTRAL) (the Cochrane Library 2014, issue 7)
-
MEDLINE (accessed through PubMed) (1966 to 1 August 2014)
-
EMBASE (accessed through Ovid) (1966 to 1 August 2014)
The search strings used are reported in Appendix 2.
The strategy was combined with a validated filter to retrieve clinical trials (see section 6.4.11 of the Cochrane Handbook for Systematic Reviews of Interventions (Lefebvre 2011)).
In addition to the search strategy, we handsearched references from relevant studies to identify any further studies. We contacted principal authors to identify any further studies or data that may be relevant to this review.
Data collection and analysis
Two authors (IA and SC) assessed eligibility of articles retrieved via the search strategy from the title and abstract, with differences resolved through the input of a third author (JW). When there was insufficient information for assessment, the authors reviewed the full articles.
Two authors (IA and SC) independently evaluated all RCTs found in the search in order to rule out those that did not meet the inclusion criteria. We evaluated those studies for probable inclusion by critical reading of the whole article. There was no blinding of the author as to the origin or conclusions of the article for eligibility assessment, data extraction or quality assessment.
If necessary, we sought information from the principal investigator of the trial concerned. Two authors (IA and SC) independently extracted the data to ensure validity, and we resolved any discrepancies by an open discussion between all investigators (IA, SC, JW, KF).
To evaluate the methodological quality of selected studies, the authors independently assessed the studies with respect to the criteria set out in the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0.(Higgins 2011a)
Each author independently undertook the extraction and registration of data from each study using data and study forms specifically designed for this review. The data extracted included details of the methodology used, the characteristics of the study participants, the type of interventions undertaken, the comparison groups, and the results obtained including the follow‐up period.
We resolved any disagreement by consensus, or with the input of a third member (JW) of the review team.
We analysed all patients initially randomised on an intention‐to‐treat (ITT) basis. The ITT population was defined as randomised patients, irrespective of whether they received any form of treatment. For dichotomous variables, we calculated risk ratio (RR) with a 95% confidence interval (CI). We used a random‐effects model for the pooled analysis. We undertook the meta‐analysis in Review Manager 5.3 (RevMan 2014).
We also calculated the proportion of people alive at 6, 12 and 24 months from both treatment groups. This data was used to generate survival curves for each of the subgroups.
Subgroup analysis
We performed subgroup analyses for the outcomes of survival (at 6 months, 12 months and 24 months) and tumour response. We categorised data from included studies into the subgroups:
-
'undifferentiated' ‐ if the study did not differentiate between patients with limited disease and extensive disease small cell lung cancer;
-
LD‐SCLC ‐ if the study presented data specifically from patients with limited disease small cell lung cancer;
-
ED‐SCLC ‐ if the study presented data specifically from patients with extensive disease small cell lung cancer.
This was undertaken in order to determine if there are differences between the treatment groups depending on the stage of the SCLC.
Heterogeneity assessment
Where substantial heterogeneity occurred, we explored the effect of potential sources of heterogeneity in an attempt to identify the cause of the heterogeneity. Substantial heterogeneity was considered to exist when the I2 value was greater than 50% (Higgins 2003).
We identified the following potential sources of heterogeneity (postulated a priori).
-
Quality of studies. We deemed studies to be of higher quality if they satisfied the following criteria in the risk of bias assessment:
-
-
incomplete outcome data addressed for all assessed outcomes;
-
free of selective reporting;
-
free of other bias.
-
(Please refer to the 'Risk of bias' table for quality assessment).
In the current update, the studies were also assessed for quality using the GRADE approach as described in the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0(Schünemann 2011). Using this method, studies were given a quality rating of High, Moderate, Low or Very Low.
-
Full article publication versus abstract publication.
-
Use of radiotherapy as a co‐intervention.
Sensitivity analysis
We performed a sensitivity analysis, systematically excluding studies from the overall analysis based on the potential sources of heterogeneity hypothesised above. This included conducting a sensitivity analysis involving only studies that used radiotherapy treatment in order to explore the potential influence of platinum‐based chemotherapy regimens and the use of radiotherapy on toxicity.
Results
Description of studies
Search results
The search strategies used for this review are outlined in Appendix 1 (strategy for the original review) and Appendix 2 (strategy for the update). The search yielded 3669 search results. We excluded 112 results as they were duplicates. A further 3525 records were excluded by abstract as they did not meet the inclusion criteria. This left 32 studies to be included in the review, including 29 from the original review (Figure 1).
Study flow diagram
Study characteristics
The 32 studies included 6075 participants who were randomised. Of these, 3036 were assigned to platinum‐based chemotherapy and 3039 to non‐platinum‐based chemotherapy regimens. All studies reported ITT analyses, thus recomputation of outcomes was not required.
Of the included studies, nineteen were conducted in centres throughout Europe, nine in North America, three in Asia and one (Quoix 2005) conducted in both Canada and Europe. All studies were one or more years in duration. The shortest study, Wolf 1987, lasted 12 months and the longest study, Urban 1999a, lasted six years.
A summary of the included studies is presented in Table 1. A detailed description of each study is presented in the section 'Characteristics of included studies'.
| Study ID | Number of Patients | Radiotherapy | |||
| First Author | Year | Platinum group | Non‐platinum group | Total | |
| Baka | 2008 | 141 | 139 | 280 | PCI, TRT |
| Chahinian | 1989 | 105 | 103 | 208 | PCI |
| Creech | 1982 | 21 | 19 | 40 | ‐ |
| de Jong | 2007 | 101 | 102 | 203 | ‐ |
| Eagan | 1981 | 31 | 31 | 62 | TRT |
| Evans | 1987 | 145 | 144 | 289 | PCI, TRT |
| Farris | 1993 | 56 | 57 | 113 | PCI, TRT |
| Fukuoka | 1986 | 35 | 34 | 69 | TRT |
| Fukuoka | 1991 | 97 | 97 | 194 | TRT |
| Gatzemeier | 1994 | 171 | 173 | 344 | ‐ |
| Goodman | 1990 | 194 | 194 | 388 | PCI, TRT |
| Greco | 2005 | 60 | 60 | 120 | PCI |
| Havemann | 1987 | 150 | 152 | 302 | PCI, TRT |
| Jones | 1993 | 54 | 50 | 104 | PCI, TRT |
| Kanitz | 1992 | 59 | 52 | 111 | ‐ |
| Lyss | 2002 | 12 | 13 | 25 | PCI |
| Postmus | 1992 | 60 | 63 | 123 | PCI |
| Postmus | 1996 | 70 | 73 | 143 | ‐ |
| Quoix | 2005 | 41 | 41 | 82 | ‐ |
| Roth | 1992 | 148 | 146 | 294 | PCI |
| Sculier | 1990 | 95 | 106 | 201 | PCI |
| Sculier | 1993 | 107 | 108 | 215 | PCI, TRT |
| Sekine | 2014 | 30 | 32 | 62 | ‐ |
| Smith | 1991 | 47 | 48 | 95 | PCI |
| Souhami | 1997 | 80 | 75 | 155 | ‐ |
| Sundstrom | 2002 | 218 | 218 | 436 | PCI, TRT |
| Urban | 1999a | 191 | 203 | 394 | PCI, TRT |
| Urban | 1999b | 229 | 228 | 457 | PCI, TRT |
| Veronesi | 1994 | 70 | 66 | 136 | PCI, TRT |
| Wampler | 1991 | 85 | 85 | 170 | ‐ |
| White | 2001 | 60 | 59 | 119 | PCI, TRT |
| Wolf | 1987 | 73 | 68 | 141 | PCI, TRT |
| Total: 32 | ‐ | 3036 | 3039 | 6075 | ‐ |
PCI ‐ Prophylactic Cranial Irradiation
TRT ‐ Thoracic Radiotherapy
Study chemotherapy interventions
The most common platinum agent was cisplatin, used in 24 studies (75%). Carboplatin was used in the remaining eight studies (25%). The most common non‐platinum agents used were etoposide (26 trials, 81.25%), cyclophosphamide (24 trials, 75%), doxorubicin (20 trials, 62.5%) and vincristine (18 trials, 56.25%).
Radiotherapy as a co‐intervention
Of the 32 included studies, seven used only prophylactic cranial irradiation (PCI), three used only thoracic radiotherapy (TRT), and thirteen used both. The remaining nine studies did not involve any form of radiotherapy. The use of radiotherapy by study is presented in Table 1.
Subgroups
The 32 studies were also divided a priori into subgroup comparisons according to disease staging, as indicated in Table 2.
| Undifferentiated | LD‐SCLC | ED‐SCLC | |||
| Creech | 1982 | Eagan | 1981 | Chahinian | 1989 |
| Farris^ | 1993 | Fukuoka | 1986 | Fukuoka | 1986 |
| Fukuoka^ | 1991 | Goodman | 1990 | Gatzemeier | 1994 |
| Postmus | 1992 | Havemann | 1987 | Greco | 2005 |
| Sculier | 1990 | Jones | 1993 | Havemann | 1987 |
| Sculier | 1993 | Sundstrom | 2002 | Jones | 1993 |
| Smith | 1991 | Urban* | 1999a | Kanitz | 1992 |
| Souhami | 1997 | Wolf | 1987 | Lyss | 2002 |
| Urban | 1999b | Baka* | 2008 | Postmus | 1996 |
| Veronesi | 1994 | ‐ | ‐ | Quoix | 2005 |
| White | 2001 | ‐ | ‐ | Roth | 1992 |
| ‐ | ‐ | ‐ | ‐ | Sundstrom | 2002 |
| ‐ | ‐ | ‐ | ‐ | Urban* | 1999a |
| ‐ | ‐ | ‐ | ‐ | Wampler | 1991 |
| ‐ | ‐ | ‐ | ‐ | Wolf | 1987 |
| ‐ | ‐ | ‐ | ‐ | de Jong | 2007 |
| ‐ | ‐ | ‐ | ‐ | Sekine | 2014 |
| Baka* | 2008 | ||||
^Presented undifferentiated survival data; response data sorted by stage.
*Presented undifferentiated response data; survival data sorted by stage.
Risk of bias in included studies
We assessed study quality according to the criteria set out in the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (Higgins 2011b)). The individual assessment for each study is set out in the 'Risk of bias' graph (Figure 2).
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Twenty‐six of the 32 studies (89.66%) did not report the randomisation process clearly. Hence, they were given an 'Unclear' rating for both sequence generation and allocation concealment. Baka 2008; de Jong 2007; Havemann 1987; Sekine 2014; Urban 1999a; and Urban 1999b, were the studies that did describe the randomisation process. Of these, all adequately described sequence generation, but not allocation concealment.
No study described the use of blinding for any participants, investigators or any of the outcomes.
Seventeen of the 32 studies (53.13%) addressed the issue of incomplete survival data. The same seventeen studies also addressed the issue of incomplete tumour response data ‐ the only exception was Sundstrom 2002 which did not report tumour response data as an outcome. Fourteen of the 32 studies (43.75%) addressed the issue of incomplete toxicity data. No study addressed the issue of incomplete quality of life data.
Fourteen of the 32 studies (43.75%) were judged to be free of selective reporting. Creech 1982 was given a rating of 'high risk' for this criterion because it did not report survival data, an outcome that would be expected from studies in this area. The remaining 17 studies were given a rating of 'Unclear'.
All studies were deemed to be free from other bias, except Souhami 1997. This is due to the skewed treatment protocol used in the study ‐ the non‐platinum treatment arm consisting of a single oral agent compared with the platinum treatment arm consisting of multiple intravenous agents.
It was noted that one study, Sekine 2014, had to be terminated early due to treatment‐related toxicity in the non‐platinum arm.
Effects of interventions
Survival
Survival at six months
Thirty studies involving 5755 participants were included in the six‐month survival analysis. The included studies that did not present survival data were Creech 1982 and Baka 2008. Of the participants, 2874 of 5755 received a platinum‐based chemotherapy regimen and the remaining 2881 received a non‐platinum‐based chemotherapy regimen. At six months, 4066 participants were alive: 2083 from the platinum‐based groups and 1983 from the non‐platinum‐based treatment groups. There was no statistically significant difference between interventions (risk ratio (RR) 1.04, 95% confidence interval (CI) 1.00 to 1.09). There was no substantial heterogeneity present in the data (I2 = 44%).
Subgroup: 'undifferentiated'
Ten studies reported data from six‐month survival comparisons, but did not differentiate between limited and extensive disease. This included 1808 participants, 901 receiving a platinum‐based and 907 receiving a non‐platinum‐based regimen. At six months, 1231 participants were alive: 624 from the platinum‐based arm and 607 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 1.02, 95% CI 0.94 to 1.10). There was no substantial heterogeneity present in the data (I2 = 36%).
Subgroup: LD‐SCLC
Eight studies reported data from six‐month survival comparisons for participants with limited disease, involving 1044 participants. Of these 516 received a platinum‐based and 528 received a non‐platinum‐based regimen. At six months, 926 participants were alive: 459 from the platinum‐based arm and 467 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 1.00, 95% CI 0.94 to 1.07). There was no substantial heterogeneity present in the data (I2 = 45%).
Subgroup: ED‐SCLC
Eighteen studies reported data from six‐month survival comparisons for participants with extensive disease, involving 2903 participants. Of these, 1457 received a platinum‐based and 1446 received a non‐platinum‐based regimen. At six months, 1909 participants were alive: 1000 from the platinum‐based arm and 909 from the non‐platinum‐based arm. In contrast to the other subgroups above, there was a statistically significant effect favouring platinum‐based regimens (RR 1.09, 95% CI 1.02 to 1.17). There was no substantial heterogeneity present in the data (I2 = 37%).
Survival at 12 months
31 studies involving 6034 participants were included in the 12‐month survival analysis. The only included study that did not present survival data was Creech 1982. Of these participants, 3015 received a platinum‐based chemotherapy regimen and 3019 received a non‐platinum‐based chemotherapy regimen. At 12 months, 2146 participants were alive: 1100 from the platinum‐based groups and 1046 from the non‐platinum‐based treatment groups. Similar to six‐month survival, at 12 months there was no statistically significant difference between interventions (RR 1.06, 95% CI 0.97 to 1.16). There was no substantial heterogeneity present in the data (I2 = 37%).
Subgroup: 'undifferentiated'
Ten studies reported data from 12‐month survival comparisons, but did not differentiate between limited and extensive disease. This included 1808 participants, 901 receiving a platinum‐based and 907 receiving a non‐platinum‐based regimen. At 12 months, 559 participants were alive: 278 from the platinum‐based arm and 281 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 1.00, 95% CI 0.82 to 1.22). There was no substantial heterogeneity present in the data (I2 = 46%).
Subgroup: LD‐SCLC
Nine studies reported data from 12‐month survival comparisons for participants with limited disease, involving 1209 participants. Of these, 597 received a platinum‐based and 612 received a non‐platinum‐based regimen. At 12 months, 701 participants were alive: 363 from the platinum‐based arm and 338 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 1.15, 95% CI 0.95 to 1.39). There was substantial heterogeneity present in the data (I2 = 64%).
Subgroup: ED‐SCLC
Nineteen studies reported data from 12‐month survival comparisons for participants with extensive disease, involving 3017 participants. Of these, 1517 received a platinum‐based and 1500 received a non‐platinum‐based regimen. At 12 months, 886 participants were alive: 459 from the platinum‐based arms and 427 from the non‐platinum‐based arm. In contrast to the corresponding subgroup at six months, there was no statistically significant difference between interventions (RR 1.06, 95% CI 0.94 to 1.18). There was no substantial heterogeneity present in the data (I2 = 4%).
Survival at 24 months
Twenty seven studies involving 5398 participants were included in the 24‐month survival analysis. The four studies that were included in the 12‐month survival analyses, but that did not report 24‐month survival data were Evans 1987, Quoix 2005, de Jong 2007 and Sekine 2014. All were part of the ED‐SCLC subgroup. The remaining 27 studies were the same as those included for 12‐month survival. Of the participants, 2698 of 5398 received a platinum‐based chemotherapy regimen and the remaining 2700 received a non‐platinum‐based chemotherapy regimen. At 24 months, 566 participants were alive: 287 from the platinum‐based groups and 279 from the non‐platinum‐based treatment groups. As with 12‐month survival, there was no statistically significant difference between interventions (RR 1.06, 95% CI 0.85 to 1.31). There was no substantial heterogeneity present in the data (I2 = 31%).
Subgroup: 'undifferentiated'
Ten studies reported data from 24‐month survival comparisons, but did not differentiate between limited and extensive disease. This included 1808 participants, 901 receiving a platinum‐based and 907 receiving a non‐platinum‐based regimen. At 24 months, 146 participants were alive: 70 from the platinum‐based arm and 76 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 0.97, 95% CI 0.71 to 1.33). There was no heterogeneity (I2 = 0%).
Subgroup: LD‐SCLC
Nine studies reported data from 12‐month survival comparisons for participants with limited disease, involving 1209 participants. Of these, 597 received a platinum‐based and 612 received a non‐platinum‐based regimen. At 24 months, 255 participants were alive: 133 from the platinum‐based arm and 122 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 1.07, 95% CI 0.7 to 1.65). There was substantial heterogeneity present in the data (I2 = 57%).
Subgroup: ED‐SCLC
Fifteen studies reported data from 24‐month survival comparisons for participants with extensive disease, involving 2381 participants. Of these, 1200 received a platinum‐based and 1181 received a non‐platinum‐based regimen. At 24 months, 165 participants were alive: 84 from the platinum‐based arms and 81 from the non‐platinum‐based arm. There was no statistically significant difference between interventions (RR 1.11, 95% CI 0.71 to 1.75). There was substantial heterogeneity present in the data (I2 = 35%).
Survival by subgroup
The proportion of people alive at 6, 12 and 24 months from both treatment groups is presented in Table 3. Figure 3 illustrates the survival curves data based on the subgroup data.
Survival by Subgroup
| Subgroup | % Survival | ||
| 6 months | 12 months | 24 months | |
| Undifferentiated | 68.09% | 30.92% | 8.08% |
| LD‐SCLC | 88.70% | 57.98% | 21.09% |
| ED‐SCLC | 65.36% | 29.37% | 6.93% |
Tumour response
Overall response
Thirty‐one studies involving 5651 participants were included in the overall response analysis. Of the 31 studies, 30 were the same as for 12‐month survival. The difference was due to the exclusion of Sundstrom 2002 (presented survival data but not response data) and the inclusion of Creech 1982 (presented response data but not survival data). Of the participants, 2824 of 5651 received a platinum‐based chemotherapy regimen and the remaining 2827 received a non‐platinum‐based chemotherapy regimen. The number of participants with a recorded overall response was 3562, consisting of 1846 participants from the platinum arm and 1716 participants from the non‐platinum arm. There was no statistically significant difference between interventions (RR 1.06, 95% CI 0.98 to 1.13). There was substantial heterogeneity present in the data (I2 = 66%).
Subgroup: 'undifferentiated'
Eleven studies reported data from overall response comparisons, but did not differentiate between limited and extensive disease. This included 2225 participants, 1105 receiving a platinum‐based and 1120 receiving a non‐platinum‐based regimen. A total of 1339 participants recorded an overall response, consisting of 682 participants from the platinum arm and 657 participants from the non‐platinum arm. There was no statistically significant difference between interventions (RR 1.04, 95% CI 0.89 to 1.21). There was substantial heterogeneity present in the data (I2 = 77%).
Subgroup: LD‐SCLC
Eight studies reported data from overall response comparisons for limited disease. This included 809 participants, 402 receiving a platinum‐based and 407 receiving a non‐platinum‐based regimen. A total of 613 participants recorded an overall response, consisting of 312 participants from the platinum arm and 301 participants from the non‐platinum arm. There was no statistically significant difference between interventions (RR 1.03, 95% CI 0.94 to 1.12). There was no substantial heterogeneity present in the data (I2 = 23%).
Subgroup: ED‐SCLC
Eighteen studies reported data from overall response comparisons for participants with ED‐SCLC. This subgroup consisted of 2617 participants, with 1317 receiving a platinum‐based chemotherapy regimen and 1300 receiving a non‐platinum‐based chemotherapy regimen. A total of 1610 participants recorded an overall response, consisting of 852 participants from platinum arms and 758 participants from non‐platinum arms. Similar to other subgroups above, there was no statistically significant difference between interventions (RR 1.07, 95% CI 0.96 to 1.19). There was substantial heterogeneity present in the data (I2 = 66%).
Complete response
Thirty studies involving 5599 participants were included in the complete response analysis. Creech 1982 (included in overall response) was excluded because it did not present complete response data. Of the 5599 participants, 2799 received a platinum‐based chemotherapy regimen and the remaining 2800 participants received a non‐platinum‐based chemotherapy regimen. Of the participants, 1243 recorded a complete response, consisting of 701 participants from the platinum arm and 542 participants from the non‐platinum arm. There was a statistically significant effect favouring platinum‐based chemotherapy regimens (RR 1.32, 95% CI 1.14 to 1.54). There was no substantial heterogeneity present in the data (I2 = 46%).
Subgroup: 'Undifferentiated'
Ten studies reported data from complete response comparisons, but did not differentiate between limited and extensive disease. This included 2175 participants, 1080 receiving a platinum‐based and 1095 receiving a non‐platinum‐based regimen. A total of 427 participants recorded a complete response, consisting of 252 participants from the platinum arm and 212 participants from the non‐platinum arm. In contrast to the other subgroups below, there was no statistically significant difference between interventions (RR 1.26, 95% CI 0.90 to 1.77). There was substantial heterogeneity present in the data (I2 = 72%).
Subgroup: LD‐SCLC
Eight studies reported data from complete response comparisons for limited disease. This included 809 participants, 402 receiving a platinum‐based and 407 receiving a non‐platinum‐based regimen. A total of 332 participants recorded a complete response, consisting of 181 participants from the platinum arm and 151 participants from the non‐platinum arm. There was a statistically significant effect favouring platinum‐based regimens (RR 1.19, 95% CI 1.02 to 1.40). There was no heterogeneity (I2 = 0%).
Subgroup: ED‐SCLC
Eighteen studies reported data from complete response comparisons for extensive disease. This subgroup consisted of 2615 participants, with 1317 receiving a platinum‐based chemotherapy regimen and 1298 receiving a non‐platinum‐based chemotherapy regimen. A total of 447 participants recorded a complete response, consisting of 268 participants from platinum arms and 179 participants from non‐platinum arms. There was a statistically significant effect, favouring platinum‐based chemotherapy regimens (RR 1.45, 95% CI 1.17 to 1.80). There was no substantial heterogeneity present in the data (I2 = 24%).
Toxicity
Toxic death
Twenty studies reported data regarding toxic death, involving 3696 (60.83%) participants. Overall, 1840 of these participants were from platinum‐based chemotherapy groups and 1856 from non‐platinum‐based chemotherapy groups. Of these 183 died from a toxicity‐related cause, 98 from platinum‐based groups and 85 from non‐platinum‐based groups. There was no statistically significant difference between interventions (RR 1.08, 95% CI 0.72 to 1.62). There was no substantial heterogeneity present in the data (I2 = 35%).
Nausea and vomiting
Twenty‐two studies reported toxicity data regarding nausea and vomiting, involving 3961 (65.20%) participants. Of those, 1978 were from platinum‐based chemotherapy groups and 1983 from non‐platinum‐based chemotherapy groups. Overall, 721 participants experienced nausea and vomiting toxicity, 430 from platinum‐based groups and 291 from non‐platinum‐based groups. There was a statistically significant difference between interventions, with higher rates of nausea and vomiting toxicity in platinum‐based chemotherapy regimens (RR 1.52, 95% CI 1.22 to 1.88). There was substantial heterogeneity present in the data (I2 = 57%).
Alopecia
Eight studies reported toxicity data regarding alopecia, involving 1526 (25.12%) participants. Of those, 753 were from platinum‐based chemotherapy groups and 773 from non‐platinum‐based chemotherapy groups. Overall, 884 participants experienced alopecia toxicity, 470 from platinum‐based groups and 414 from non‐platinum‐based groups. There was no statistically significant difference between interventions (RR 1.14, 95% CI 1.00 to 1.30). There was substantial heterogeneity present in the data (I2 = 58%).
Infection
Twelve studies reported toxicity data regarding infection, involving 2184 (35.95%) participants. Of those, 1096 were from platinum‐based chemotherapy groups and 1088 from non‐platinum‐based chemotherapy groups. Overall, 293 participants experienced infections, 98 from platinum‐based groups and 195 from non‐platinum‐based groups. There was no statistically significant difference between interventions (RR 0.69 95% CI 0.43 to 1.09). There was substantial heterogeneity present in the data (I2 = 60%).
Anaemia
Twelve studies reported toxicity data regarding anaemia, involving 2219 (36.53%) participants. Of these, 1112 were from platinum‐based chemotherapy groups and 1107 from non‐platinum‐based chemotherapy groups. Overall, 316 participants experienced anaemia toxicity, 169 from platinum‐based groups and 147 from non‐platinum‐based groups. There was no statistically significant difference between interventions (RR 1.19, 95% CI 0.86 to 1.67). There was substantial heterogeneity present in the data (I2 = 52%).
Leukopenia
Seventeen studies reported toxicity data regarding leukopenia, involving 2556 (42.07%) participants. Of these, 1277 were from platinum‐based chemotherapy groups and 1279 from non‐platinum‐based chemotherapy groups. Overall, 1175 participants experienced leukopenia toxicity, 526 from platinum‐based groups and 649 from non‐platinum‐based groups. There was no statistically significant difference between interventions (RR 0.84, 95% CI 0.67 to 1.05). There was substantial heterogeneity present in the data (I2 = 87%)
Thrombocytopenia
Nineteen studies reported toxicity data regarding thrombocytopenia, involving 3276 (53.93%) participants. Of those, 1637 were from platinum‐based chemotherapy groups and 1639 from non‐platinum‐based chemotherapy groups. Overall, 467 participants experienced thrombocytopenia toxicity, 305 from platinum‐based groups and 162 from non‐platinum‐based groups. There was a statistically significant difference between interventions, with higher rates of thrombocytopenia toxicity in platinum‐based chemotherapy regimens (RR 1.89, 95% CI 1.37 to 2.61). There was substantial heterogeneity present in the data (I2 = 62%).
Granulocytopenia
Five studies reported toxicity data regarding granulocytopenia, involving 1229 (22.22%) participants. Of those, 622 were from platinum‐based chemotherapy groups and 607 from non‐platinum‐based chemotherapy groups. Overall, 517 participants experienced granulocytopenia toxicity, 244 from platinum‐based groups and 273 from non‐platinum‐based groups. There was no statistically significant difference between interventions (RR 0.89, 95% CI 0.79 to 1.00). There was no heterogeneity (I2 = 0%).
Quality of life
Data about quality of life (QOL) were available from four studies. It was not possible to combine these in a meta‐analysis because none of the studies reported enough data to calculate standardised mean differences. Results from these studies are discussed individually.
Souhami 1997 measured QOL via two scales: the Rotterdam Symptom Checklist and the daily diary card. The Rotterdam Symptom Checklist measured physical symptoms, lung cancer symptoms, treatment symptoms, physical activity, psychological well‐being and general QOL. It was completed by participants at each treatment cycle and at the first two follow‐up visits. The study reported that there was no significant difference in physical symptoms, psychological well‐being or physical activity, between interventions. Treatment‐related symptoms were significantly worse, but lung cancer symptoms and general QOL were significantly better in the platinum treatment arm. Souhami 1997 was a unique study in that it was the only study to have a non‐intravenous form of delivery: its non‐platinum treatment arm (oral etoposide).
A daily diary card was used to measure acute chemotherapy‐related symptoms. It measured nausea, vomiting, appetite, pain, sleep, mood, general well‐being and physical activity on a daily basis. The findings from this study showed that there was no significant difference in vomiting or physical activity between the two treatment arms. Nausea was found to be worse in the platinum‐based treatment arm. But in all other categories, the non‐platinum‐based treatment arms were significantly worse (Souhami 1997).
Sundstrom 2002 measured QOL via the European Organisation for the Research and Treatment of Cancer (EORTC) quality of life questionnaire C30 (QLQ‐C30). The QLQ‐C30 questionnaire reports five functional scales (physical, role, emotional, cognitive and social functioning), three symptom scales (fatigue, nausea/vomiting and pain), a global health and overall QOL scale and five single item scales (dyspnoea, appetite loss, sleep disturbance, constipation and diarrhoea, and the financial impact of disease and treatment). Results from the questionnaire showed no statistically significant difference between the platinum group and the non‐platinum group for most of the scales.
Quoix 2005 measured QOL using the functional assessment of cancer therapy‐lung (FACT‐L) QOL questionnaire and a separate symptom questionnaire. The questionnaires were completed by participants at baseline and prior to each course of treatment.
The FACT‐L questionnaire assessed physical well‐being, social/family well‐being, emotional well‐being and additional concerns on a five‐point scale. These were used to determine a FACT‐L score, and the scores for physical and functional well‐being were summed to give a trial outcome index (TOI) score. Both the platinum and non‐platinum arm had improvements in FACT‐L and TOI scores compared to baseline scores following six cycles of treatments. However there was no significant difference in FACT‐L or TOI scores between treatment arms (Quoix 2005).
The symptom questionnaire employed by Quoix 2005 assessed shortness of breath, cough, chest pain, haemoptysis, anorexia, insomnia, hoarseness, fatigue and interference with daily activity. The results found both interventions improved all symptoms except haemoptysis, in more than 20% of participants when compared to baseline. However, there was no significant difference between treatments (Quoix 2005).
Sekine 2014 used two scales to assess QOL ‐ lung cancer subscale (LCS) of FACT‐L questionnaire and Euro‐QOL 5‐dimension (EQ‐5D) utility index. Assessments were performed at 3 weeks, 3 months, 6 months and 12 months after the commencement of chemotherapy. The LCS measured a number of symptoms including breathlessness, weight loss, cognitive function, appetite and chest discomfort (Cella 2002). For both the platinum group and the non‐platinum group, there were no significant differences in LCS scores at all of the time points (P values: 0.171 at 3 weeks; 0.08 at 3 months; 0.112 at 6 months; 0.371 at 12 months). The EQ‐5D measured five socially relevant domains ‐ mobility, self‐care, usual activities, pain/discomfort and anxiety/depression (Gusi 2010). For both the platinum group and the non‐platinum group, there were no significant differences in LCS scores at all of the time points (P values: 0.171 at 3 weeks; 0.08 at 3 months; 0.112 at 6 months; 0.371 at 12 months). For both the platinum group and the non‐platinum group, there were no significant differences in EQ‐5D utility index scores at all of the time points (P values: 0.171 at 3 weeks; 0.08 at 3 months; 0.112 at 6 months; 0.371 at 12 months).
Sensitivity analysis
Quality of studies
The effect of the quality of studies was examined by including all studies that were deemed to be of high quality using the GRADE assessment approach as discussed in the Methods. As such, the included studies for the sensitivity analysis of quality were: Baka 2008; de Jong 2007; Evans 1987; Fukuoka 1991; Havemann 1987; Kanitz 1992; Lyss 2002; Postmus 1996; Roth 1992; Sculier 1990; Sculier 1993; Smith 1991; Sundstrom 2002; Urban 1999a; Veronesi 1994 and White 2001, . The results are presented in Table 4.
| Outcome | All studies | Higher quality studies |
| 6‐month survival | NSSD | NSSD |
| 12‐month survival | NSSD | NSSD |
| 24‐month survival | NSSD | NSSD |
| Overall response | NSSD | P |
| Complete response | P | P |
| Toxic death | NSSD | NSSD |
| Nausea and Vomiting | NP | NP |
| Alopecia | NSSD | NP |
| Infection | NSSD | P |
| Anaemia | NSSD | NSSD |
| Leukopenia | NSSD | P |
| Thrombocytopenia | NP | NSSD |
| Granulocytopenia | NSSD | NSSD |
NSSD ‐ No statistically significant difference between treatment groups.
P ‐ Statistically significant effect favouring platinum‐based regimens.
NP ‐ Statistically significant effect favouring non‐platinum‐based regimens.
It can be seen from this table that there were five differences in outcomes between the sensitivity analysis of quality and the overall analysis. The outcomes affected were overall response, alopecia, infection, leukopenia and thrombocytopenia toxicities. For overall response and infection, there was no significant difference between the two treatment groups, however, when only the higher quality studies were included there was a significant difference favouring the platinum group. There was no significant difference between the two treatment groups in terms of alopecia toxicity. However, in the sensitivity analysis for quality, the platinum‐based treatment group had a significantly higher level of alopecia toxicity. The reverse was true for thrombocytopenia toxicity. There was a significantly higher rate of thrombocytopenia toxicity in the platinum‐based treatment group in the overall analysis, but this difference was not statistically significant in the sensitivity analysis for quality. Finally, for leukopenia toxicity, there was no significant difference between the treatment groups in the overall analysis. However, in the sensitivity analysis for quality, the rate of leukopenia toxicity was significantly higher in the non‐platinum‐based treatment group.
Full article versus abstract publication
Full articles for all studies were available, hence this sensitivity analysis was not conducted.
Use of radiotherapy
The effect of radiotherapy was explored by including all studies that used radiotherapy. Hence, the following studies were included in the analysis: Baka 2008; Chahinian 1989; Eagan 1981; Evans 1987; Farris 1993; Fukuoka 1986; Fukuoka 1991; Goodman 1990; Greco 2005; Havemann 1987; Jones 1993; Lyss 2002; Postmus 1992; Roth 1992; Sculier 1990; Sculier 1993; Smith 1991; Sundstrom 2002; Urban 1999a; Urban 1999b; Veronesi 1994; White 2001 and Wolf 1987. The results are presented in Table 5.
| Outcome | All studies | Only studies using radiotherapy |
| 6‐month survival | NSSD | NSSD |
| 12‐month survival | NSSD | NSSD |
| 24‐month survival | NSSD | NSSD |
| Overall response | NSSD | NSSD |
| Complete response | P | P |
| Toxic death | NSSD | NSSD |
| Nausea and Vomiting | NP | NP |
| Alopecia | NSSD | NSSD |
| Infection | NSSD | NSSD |
| Anaemia | NSSD | NSSD |
| Leukopenia | NSSD | NSSD |
| Thrombocytopenia | NP | NP |
| Granulocytopenia | NSSD | NSSD |
NSSD ‐ No statistically significant difference between treatment groups.
P ‐ Statistically significant effect favouring platinum‐based regimens.
NP ‐ Statistically significant effect favouring non‐platinum‐based regimens.
It can be seen from this table that there is no significant difference between the outcomes of all studies compared with the outcomes of only studies that used radiotherapy.
Heterogeneity analysis
It can be seen from the results that there was substantial heterogeneity associated with a number of outcomes, including overall response, nausea and vomiting, alopecia, leukopenia and thrombocytopenia toxicity. For all of these outcomes, possible reasons for the heterogeneity were explored. No specific causes of heterogeneity were found for any of the outcomes, except for thrombocytopenia toxicity (discussed below). Possible reasons for heterogeneity are mentioned in the Tumour Response and Toxicity sections in the Discussion. It is important to note that the results of all outcomes with substantial heterogeneity must be treated with caution.
Discussion
This systematic review included 32 trials and 6075 patients, which provides adequate data to undertake meaningful meta‐analyses.
Cisplatin and carboplatin are the only platinum‐based agents used in SCLC, with cisplatin more widely used than carboplatin. Of the non‐platinum agents, etoposide (E) and cyclophosphamide (C) are the most commonly used, followed by doxorubicin (D), adriamycin (A) and vincristine (V) usually in the triplet combination of either CEV, CDE or CAV (Abeloff 2004).
Co‐intervention with radiotherapy occurred in approximately 72% of studies indicating that this plays a major role in the treatment of SCLC. It should be noted that the sensitivity analysis to exclude any confounding effects of radiotherapy did not find any statistically significant differences in any of the outcomes, including toxicity, from the overall analysis.
Survival
No statistically significant difference between platinum‐based chemotherapy regimens and non‐platinum‐based chemotherapy regimens was found for survival at 6 months, 12 months or 24 months.
Subgroup analyses demonstrated that survival for the 'undifferentiated' subgroup was similar to that of extensive disease. This is explained by the high proportion of extensive disease‐stage patients in the studies in the 'undifferentiated' group. For example, Urban 1999a had 360 extensive disease‐stage patients and 97 limited disease‐stage patients. The high proportion of extensive disease‐stage participants reflects the fact that SCLC is an aggressive disease that metastasises early and usually presents with extensive disease (Abeloff 2004).
Despite survival being better in limited disease, the slopes of the survival curves (in Figure 3) for all three subgroups are relatively similar. This implies that the effect of chemotherapy does not differ depending on the stage of the disease and survival can be attributed to the disease extent at diagnosis (ACN 2004).
The outcomes show that even with treatment, long‐term survival with SCLC is uncommon; the highest proportions of survivors at 24 months are recorded in the limited disease group (21.09%).
Tumour response
There was no statistically significant difference between platinum‐based chemotherapy regimens and non‐platinum‐based chemotherapy regimens in terms of overall response when the results of all studies were considered. However, the sensitivity analysis revealed that, when only studies of higher quality were included, there was a statistically significant difference in response rates favouring platinum‐based chemotherapy regimens. In addition, platinum‐based chemotherapy regimens did demonstrate a significantly higher complete response than non‐platinum‐based chemotherapy regimens. However this does not translate into improved survival in the short, intermediate or long‐term. These findings are consistent with the knowledge that SCLC is initially responsive to chemotherapy, but relapse is common and further response to chemotherapy is poor (Abeloff 2004; Peckham 1995). This is thought to be due to resistant subpopulations of tumour cells developing over time (Evans 1986).
It should be noted that there was substantial heterogeneity in the overall response analysis. A definite reason for this could not be identified. However, a possible explanation for this is the variation in staging and response assessment methods used. Since this review involved some studies that were conducted decades apart, earlier studies may have used less sensitive staging and response assessment methods than later studies. Similarly, some study centres may not have had the sophisticated resources available for staging and response assessment as other study centres. So while the definition of overall response has remained constant, the assessment methods used may have varied. As such, this variation may have contributed to the heterogeneity in the overall response analysis.
Toxicity
"Given that there was no significant difference in survival between the two types of regimen, the balance between the benefits and risks of treatment is more important. However the adverse events were not well‐reported. This systematic review has shown that platinum‐based regimens result in higher rates of nausea and vomiting, and thrombocytopenia than non‐platinum‐based regimens. There was no statistically significant difference between the two groups with respect to toxic death, alopecia, infection, anaemia, leukopenia and granulocytopenia. Therefore platinum‐based chemotherapy regimens did not result in statistically significant lower rates of toxicity in any of the toxicity analyses.
It should also be noted that there was substantial heterogeneity in many of the toxicity analyses. The cause of this heterogeneity was difficult to identify. This is surprising considering the fact that all studies adhered to the grade 3/4 WHO definitions for each of the toxicity outcomes. One possible reason for the heterogeneity may be differences in the therapies used to prevent or treat toxicities, including antiemetics, antibiotics, cytokines, blood and platelet transfusions and granulocyte‐colony stimulating factor (G‐CSF). In particular, it should be noted that 5‐HT3 antagonists, which are powerful anti‐emetic agents, were only discovered in 1988 and did not come into widespread use until a decade later (Stubblefield 2009) – after the publication date of the majority of studies in this review. This could have resulted in higher grade 3/4 nausea and vomiting toxicity with platinum‐based chemotherapy regimens, than would be seen today with the availability of a wider range of more effective anti‐emetic agents, including 5‐HT3 antagonists.
Quality of life
A major limitation of this systematic review was the insufficient data to conduct a meta‐analysis on quality of life (QOL). A qualitative description of the results from four studies that reported QOL did not indicate substantial differences between platinum‐ and non‐platinum‐based treatment regimens.
While it has been shown by a number of studies that tumour progression is associated with a poorer QOL (Gralla 2004), it is not reasonable to infer that a treatment that produces a tumour response will necessarily improve QOL. While platinum‐based regimens, with the better tumour response may improve QOL through tumour control, this is offset by the higher incidence of some toxicities. Therefore, the only reliable method of determining which treatment regimen is associated with a better QOL is for more RCTs to include QOL in their outcome assessment. This systematic review highlights the current lack of QOL data and the need for future studies to incorporate QOL as an outcome measure.
Limitations
There are a number of limitations associated with this systematic review. Firstly, a systematic review should ideally be conducted using individual patient data (IPD). However, this is rarely done in practice, as it is not considered practical to do so in a large number of cases, largely because IPD from studies are not always easily obtainable (Higgins 2011c. For this systematic review, IPD were not available for the studies, despite written requests to the authors. While this is understandable, considering a significant number of the included studies were conducted more than a decade earlier, it does limit the quality and reliability of this systematic review.
Secondly, the effect of including studies spanning a large time period is not known, and some chemotherapy regimens used in earlier studies may now be considered out of date. This is particularly the case with some non‐platinum‐based regimens. Hence, it should be kept in mind that the results from these studies may have affected the overall results. Thirdly, the effect of co‐intervention with radiotherapy in limited disease is unclear as none of the studies in this subgroup included radiotherapy and thus all were excluded in the sensitivity analysis for this variable.
Finally, although survival graphs were given by a majority of studies, the associated log hazard ratio (HR) and summary statistics were rarely provided. As a result, the survival data from studies could not be combined to obtain an overall HR. Instead, data on survival at pre‐defined time points (6 months, 12 months and 24 months) were used. Because of the lack of HR and summary statistic data, calculating the survival at these pre‐defined time points was considered a reasonable procedure to determine short, intermediate and long‐term survival. In addition, this method of using predefined time intervals has been used in other published systematic reviews exploring survival in SCLC (Pujol 2000).
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Comparison 1 Treatment Regimens, Outcome 1 6‐month survival.
Comparison 1 Treatment Regimens, Outcome 2 12‐month survival.
Comparison 1 Treatment Regimens, Outcome 3 24‐month survival.
Comparison 1 Treatment Regimens, Outcome 4 Overall response.
Comparison 1 Treatment Regimens, Outcome 5 Complete response.
Comparison 1 Treatment Regimens, Outcome 6 Toxic Death.
Comparison 1 Treatment Regimens, Outcome 7 Nausea and Vomiting.
Comparison 1 Treatment Regimens, Outcome 8 Alopecia.
Comparison 1 Treatment Regimens, Outcome 9 Infection.
Comparison 1 Treatment Regimens, Outcome 10 Anaemia.
Comparison 1 Treatment Regimens, Outcome 11 Leukopenia.
Comparison 1 Treatment Regimens, Outcome 12 Thrombocytopenia.
Comparison 1 Treatment Regimens, Outcome 13 Granulocytopenia.
| Study ID | Number of Patients | Radiotherapy | |||
| First Author | Year | Platinum group | Non‐platinum group | Total | |
| Baka | 2008 | 141 | 139 | 280 | PCI, TRT |
| Chahinian | 1989 | 105 | 103 | 208 | PCI |
| Creech | 1982 | 21 | 19 | 40 | ‐ |
| de Jong | 2007 | 101 | 102 | 203 | ‐ |
| Eagan | 1981 | 31 | 31 | 62 | TRT |
| Evans | 1987 | 145 | 144 | 289 | PCI, TRT |
| Farris | 1993 | 56 | 57 | 113 | PCI, TRT |
| Fukuoka | 1986 | 35 | 34 | 69 | TRT |
| Fukuoka | 1991 | 97 | 97 | 194 | TRT |
| Gatzemeier | 1994 | 171 | 173 | 344 | ‐ |
| Goodman | 1990 | 194 | 194 | 388 | PCI, TRT |
| Greco | 2005 | 60 | 60 | 120 | PCI |
| Havemann | 1987 | 150 | 152 | 302 | PCI, TRT |
| Jones | 1993 | 54 | 50 | 104 | PCI, TRT |
| Kanitz | 1992 | 59 | 52 | 111 | ‐ |
| Lyss | 2002 | 12 | 13 | 25 | PCI |
| Postmus | 1992 | 60 | 63 | 123 | PCI |
| Postmus | 1996 | 70 | 73 | 143 | ‐ |
| Quoix | 2005 | 41 | 41 | 82 | ‐ |
| Roth | 1992 | 148 | 146 | 294 | PCI |
| Sculier | 1990 | 95 | 106 | 201 | PCI |
| Sculier | 1993 | 107 | 108 | 215 | PCI, TRT |
| Sekine | 2014 | 30 | 32 | 62 | ‐ |
| Smith | 1991 | 47 | 48 | 95 | PCI |
| Souhami | 1997 | 80 | 75 | 155 | ‐ |
| Sundstrom | 2002 | 218 | 218 | 436 | PCI, TRT |
| Urban | 1999a | 191 | 203 | 394 | PCI, TRT |
| Urban | 1999b | 229 | 228 | 457 | PCI, TRT |
| Veronesi | 1994 | 70 | 66 | 136 | PCI, TRT |
| Wampler | 1991 | 85 | 85 | 170 | ‐ |
| White | 2001 | 60 | 59 | 119 | PCI, TRT |
| Wolf | 1987 | 73 | 68 | 141 | PCI, TRT |
| Total: 32 | ‐ | 3036 | 3039 | 6075 | ‐ |
| PCI ‐ Prophylactic Cranial Irradiation TRT ‐ Thoracic Radiotherapy | |||||
| Undifferentiated | LD‐SCLC | ED‐SCLC | |||
| Creech | 1982 | Eagan | 1981 | Chahinian | 1989 |
| Farris^ | 1993 | Fukuoka | 1986 | Fukuoka | 1986 |
| Fukuoka^ | 1991 | Goodman | 1990 | Gatzemeier | 1994 |
| Postmus | 1992 | Havemann | 1987 | Greco | 2005 |
| Sculier | 1990 | Jones | 1993 | Havemann | 1987 |
| Sculier | 1993 | Sundstrom | 2002 | Jones | 1993 |
| Smith | 1991 | Urban* | 1999a | Kanitz | 1992 |
| Souhami | 1997 | Wolf | 1987 | Lyss | 2002 |
| Urban | 1999b | Baka* | 2008 | Postmus | 1996 |
| Veronesi | 1994 | ‐ | ‐ | Quoix | 2005 |
| White | 2001 | ‐ | ‐ | Roth | 1992 |
| ‐ | ‐ | ‐ | ‐ | Sundstrom | 2002 |
| ‐ | ‐ | ‐ | ‐ | Urban* | 1999a |
| ‐ | ‐ | ‐ | ‐ | Wampler | 1991 |
| ‐ | ‐ | ‐ | ‐ | Wolf | 1987 |
| ‐ | ‐ | ‐ | ‐ | de Jong | 2007 |
| ‐ | ‐ | ‐ | ‐ | Sekine | 2014 |
| Baka* | 2008 | ||||
| ^Presented undifferentiated survival data; response data sorted by stage. | |||||
| Subgroup | % Survival | ||
| 6 months | 12 months | 24 months | |
| Undifferentiated | 68.09% | 30.92% | 8.08% |
| LD‐SCLC | 88.70% | 57.98% | 21.09% |
| ED‐SCLC | 65.36% | 29.37% | 6.93% |
| Outcome | All studies | Higher quality studies |
| 6‐month survival | NSSD | NSSD |
| 12‐month survival | NSSD | NSSD |
| 24‐month survival | NSSD | NSSD |
| Overall response | NSSD | P |
| Complete response | P | P |
| Toxic death | NSSD | NSSD |
| Nausea and Vomiting | NP | NP |
| Alopecia | NSSD | NP |
| Infection | NSSD | P |
| Anaemia | NSSD | NSSD |
| Leukopenia | NSSD | P |
| Thrombocytopenia | NP | NSSD |
| Granulocytopenia | NSSD | NSSD |
| NSSD ‐ No statistically significant difference between treatment groups. P ‐ Statistically significant effect favouring platinum‐based regimens. NP ‐ Statistically significant effect favouring non‐platinum‐based regimens. | ||
| Outcome | All studies | Only studies using radiotherapy |
| 6‐month survival | NSSD | NSSD |
| 12‐month survival | NSSD | NSSD |
| 24‐month survival | NSSD | NSSD |
| Overall response | NSSD | NSSD |
| Complete response | P | P |
| Toxic death | NSSD | NSSD |
| Nausea and Vomiting | NP | NP |
| Alopecia | NSSD | NSSD |
| Infection | NSSD | NSSD |
| Anaemia | NSSD | NSSD |
| Leukopenia | NSSD | NSSD |
| Thrombocytopenia | NP | NP |
| Granulocytopenia | NSSD | NSSD |
| NSSD ‐ No statistically significant difference between treatment groups. P ‐ Statistically significant effect favouring platinum‐based regimens. NP ‐ Statistically significant effect favouring non‐platinum‐based regimens. | ||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 6‐month survival Show forest plot | 30 | 5755 | Risk Ratio (M‐H, Random, 95% CI) | 1.04 [1.00, 1.09] |
| 1.1 Undifferentiated | 10 | 1808 | Risk Ratio (M‐H, Random, 95% CI) | 1.02 [0.94, 1.10] |
| 1.2 Limited Disease | 8 | 1044 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.94, 1.07] |
| 1.3 Extensive Disease | 18 | 2903 | Risk Ratio (M‐H, Random, 95% CI) | 1.09 [1.02, 1.17] |
| 2 12‐month survival Show forest plot | 31 | 6034 | Risk Ratio (M‐H, Random, 95% CI) | 1.06 [0.97, 1.16] |
| 2.1 Undifferentiated | 10 | 1808 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.82, 1.22] |
| 2.2 Limited Disease | 9 | 1209 | Risk Ratio (M‐H, Random, 95% CI) | 1.15 [0.95, 1.39] |
| 2.3 Extensive Disease | 19 | 3017 | Risk Ratio (M‐H, Random, 95% CI) | 1.06 [0.94, 1.18] |
| 3 24‐month survival Show forest plot | 27 | 5398 | Risk Ratio (M‐H, Random, 95% CI) | 1.06 [0.85, 1.31] |
| 3.1 Undifferentiated | 10 | 1808 | Risk Ratio (M‐H, Random, 95% CI) | 0.97 [0.71, 1.33] |
| 3.2 Limited Disease | 9 | 1209 | Risk Ratio (M‐H, Random, 95% CI) | 1.07 [0.70, 1.65] |
| 3.3 Extensive Disease | 15 | 2381 | Risk Ratio (M‐H, Random, 95% CI) | 1.11 [0.71, 1.75] |
| 4 Overall response Show forest plot | 31 | 5651 | Risk Ratio (M‐H, Random, 95% CI) | 1.06 [0.98, 1.13] |
| 4.1 Undifferentiated | 11 | 2225 | Risk Ratio (M‐H, Random, 95% CI) | 1.04 [0.89, 1.21] |
| 4.2 Limited Disease | 8 | 809 | Risk Ratio (M‐H, Random, 95% CI) | 1.03 [0.94, 1.12] |
| 4.3 Extensive Disease | 18 | 2617 | Risk Ratio (M‐H, Random, 95% CI) | 1.07 [0.96, 1.19] |
| 5 Complete response Show forest plot | 30 | 5599 | Risk Ratio (M‐H, Random, 95% CI) | 1.32 [1.14, 1.54] |
| 5.1 Undifferentiated | 10 | 2175 | Risk Ratio (M‐H, Random, 95% CI) | 1.26 [0.90, 1.77] |
| 5.2 Limited Disease | 8 | 809 | Risk Ratio (M‐H, Random, 95% CI) | 1.19 [1.02, 1.40] |
| 5.3 Extensive Disease | 18 | 2615 | Risk Ratio (M‐H, Random, 95% CI) | 1.45 [1.17, 1.80] |
| 6 Toxic Death Show forest plot | 20 | 3696 | Risk Ratio (M‐H, Random, 95% CI) | 1.08 [0.72, 1.62] |
| 7 Nausea and Vomiting Show forest plot | 22 | 3961 | Risk Ratio (M‐H, Random, 95% CI) | 1.52 [1.22, 1.88] |
| 8 Alopecia Show forest plot | 8 | 1526 | Risk Ratio (M‐H, Random, 95% CI) | 1.14 [1.00, 1.30] |
| 9 Infection Show forest plot | 12 | 2184 | Risk Ratio (M‐H, Random, 95% CI) | 0.69 [0.43, 1.09] |
| 10 Anaemia Show forest plot | 12 | 2219 | Risk Ratio (M‐H, Random, 95% CI) | 1.19 [0.86, 1.67] |
| 11 Leukopenia Show forest plot | 17 | 2556 | Risk Ratio (M‐H, Random, 95% CI) | 0.84 [0.67, 1.05] |
| 12 Thrombocytopenia Show forest plot | 19 | 3276 | Risk Ratio (M‐H, Random, 95% CI) | 1.89 [1.37, 2.61] |
| 12.1 New Subgroup | 19 | 3276 | Risk Ratio (M‐H, Random, 95% CI) | 1.89 [1.37, 2.61] |
| 13 Granulocytopenia Show forest plot | 5 | 1229 | Risk Ratio (M‐H, Random, 95% CI) | 0.89 [0.79, 1.00] |
