Platinum‐containing regimens for triple‐negative metastatic breast cancer

Summary of findings 1. Platinum compared to non‐platinum regimens for metastatic triple‐negative breast cancer: OS, PFS/TTP and OTRR

Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Platinum compared to non‐platinum chemotherapy regimens for women with metastatic triple‐negative breast cancer
Patient or population: women with metastatic triple‐negative breast cancer (mTNBC) Setting: hospital Intervention: platinum Comparison: non‐platinum chemotherapy regimens
Outcomes	Risk with non‐platinum chemotherapy regimens	Risk with platinum containing regimens	Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Overall survival (OS)	1‐year risk of death		HR 0.85 (0.73 to 1.00)	958 (6)	⊕⊕⊕⊝ MODERATE³	Heterogeneity: Chi² = 5.05, df = 5 (P = 0.41); I² 1%
	510 per 1,000 ¹	455 per 1,000 (406 to 510)²
	2‐year risk of death
	711 per 1,000 ¹	652 per 1,000 (596 to 711)²
Progression‐free survival/time to progression (PFS/TTP)	1‐year risk of progression or death		HR 0.77 (0.68 to 0.88)	1077 (8)	⊕⊝⊝⊝ VERY LOW ^{4 5 6}	Heterogeneity: Chi² = 34.78, df = 7 (P < 0.0001); I² 80%
	936 per 1,000 ¹	880 per 1,000 (846 to 911)²
	2‐year risk of progression or death
	970 per 1,000 ¹	933 per 1,000 (908 to 954)²
Objective tumour response rate (OTRR) (assessable participants)	368 per 1,000 ⁷	515 per 1,000 (449 to 585)	RR 1.40 (1.22 to 1.59)	1205 (10)	⊕⊕⊝⊝ LOW ^{4 5}	Heterogeneity: Chi² = 21.44, df = 9 (P = 0.01); I² 58%
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; HR: Hazard ratio;
GRADE Working Group grades of evidence High quality (⊕⊕⊕⊕): We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality (⊕⊕⊕⊝): We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality (⊕⊕⊝⊝): Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality (⊕⊝⊝⊝ or ⊝⊝⊝⊝): We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Each ⊝ symbol represents a downgrading of the quality of evidence one level from the highest level of 'high quality (⊕⊕⊕⊕).'
¹Estimated from the average of non‐platinum group Kaplan‐Meier probabilities from the 3 highest weighted treatment‐comparisons in Analysis 1.1. ²Estimated as 1000*(1‐S(t)^HR) where S(t) is the estimated probability of survival for non‐platinum participants and HR is the pooled hazard ratio (Guyatt 1998) ³Downgraded quality of evidence one level for ’serious imprecision’ because the confidence interval for the pooled estimate is wide and crosses or nearly crosses unity. ⁴Downgraded quality of evidence one level for ’serious indirectness’ because this outcome is a surrogate endpoint of questionable validity for assessing the more important outcome of OS in the context of metastatic breast cancer (Burzykowski 2008). ⁵Downgraded quality of evidence one level for ’serious inconsistency’ because there was substantial evidence of heterogeneity. ⁶Downgraded quality of evidence one level for suspected publication bias (forest plot asymmetry). ⁷Estimated from all 10 mTNBC treatment‐comparisons in the review with OTRR results.

Summary of findings 2. Platinum‐containing regimens and toxicity profile

Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Platinum compared to non‐platinum chemotherapy regimens for treatment related death, nausea/vomiting, nephrotoxicity, anaemia, hair loss, leukopaeniaand treatment discontinuation due to adverse event
Patient or population: women with metastatic triple‐negative breast cancer (mTNBC) Setting: hospital Intervention: platinum Comparison: non‐platinum chemotherapy regimens
Outcomes	Risk with non‐platinum chemotherapy regimens	Risk with platinum containing regimens	Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Treatment‐related death (safety population)	5 per 1000 ¹	5 per 1,000 (1 to 23)	(RR 1.06, 95% CI 0.24 to 4.61)	843 (5)	⊕⊕⊝⊝ LOW ³	Heterogeneity: P = 0.69, I² 0%
Nausea/vomiting* grade 3 or 4 (safety population)	15 per 1,000 ¹	72 per 1,000 (29 to 177)	(RR 4.77, 95% CI 1.93 to 11.81)	655 (3)	⊕⊕⊝⊝ LOW ³	Heterogeneity: P = 0.32, I² 12%
Nephrotoxicity (safety population)	‐	‐	‐	‐	‐	No trials reported this outcome for mTNBC patients.
Anaemia grade 3 or 4 (safety population)	36 per 1,000 ¹	137 per 1,000 (81 to 231)	(RR 3.80, 95% CI 2.25 to 6.42)	843 (5)	⊕⊕⊝⊝ LOW ²⁴	Heterogeneity: P = 0.04, I² 65%
Hair loss (safety population)	3 per 1000 ¹	1 per 1,000 (0 to 24)	(RR 0.33, 95% CI 0.01 to 8.04)	602 (2)	⊕⊕⊝⊝ LOW ³	Heterogeneity not applicable
Leukopaenia (safety population)	155 per 1000 ¹	169 per 1000 (130 to 220)	(RR 1.09, 95% CI 0.84 to 1.42)	843 (5)	⊕⊕⊕⊝ MODERATE ²	Heterogeneity: P = 0.75, I² 0%
Treatment discontinuation due to adverse event (safety population)	93 per 1000 ¹	82 per 1000 (55 to 123)	(RR 0.88, 95% CI 0.59 to 1.32)	843 (5)	⊕⊕⊕⊝ MODERATE ²	Heterogeneity: P = 0.07, I² 57%
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality (⊕⊕⊕⊕): We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality (⊕⊕⊕⊝): We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality (⊕⊕⊝⊝): Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality (⊕⊝⊝⊝ or ⊝⊝⊝⊝): We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Each ⊝ symbol represents a downgrading of the quality of evidence one level from the highest level of 'high quality (⊕⊕⊕⊕).'
¹Estimated from all treatment‐comparisons contributing data for pooling for this outcome (including treatment‐comparisons with non‐estimable effects due to no events in either arm). ²Downgraded quality of evidence one level for 'serious imprecision' because the confidence interval for the pooled estimate is wide. ³Downgraded quality of evidence two levels for 'very serious imprecision' because the confidence interval for the pooled estimate is very wide. ⁴Downgraded quality of evidence one level for 'serious inconsistency' because there was evidence of heterogeneity across studies (P < 0.05) *data on vomiting was included if data on nausea/vomiting was reported separately

Background

Description of the condition

Breast cancer is both the most common type of cancer in women and the most common cause of cancer death in women (Ferlay 2018a). In 2018, there was an estimated 2.1 million estimated new cases and approximately 627,000 deaths from breast cancer worldwide, with an age‐standardised death rate (ASR) of 13.0 per 100,000 (Ferlay 2018a). In the same year, the disease was the most common cancer type in more than three‐quarters of countries worldwide and was the leading cause of cancer death in more than half of countries worldwide (Ferlay 2018b).

The stage of breast cancer at the time of diagnosis is an important indicator of prognosis. Once breast cancer becomes metastatic, it is not generally considered curable and most women with metastatic disease do not survive beyond five years from the time of their metastatic diagnosis (Clements 2012). Another important predictor of prognosis is the biological subtype of breast cancer. One of these subtypes, triple‐negative breast cancer (TNBC), is characterised by a lack of expression of oestrogen receptors (ER), progesterone receptors (PgR) and human epidermal receptor 2 (HER2). TNBC comprises approximately 12% to 17% of breast cancers and is associated with shorter survival and higher likelihood of recurrence (Foulkes 2010). The median survival time for women diagnosed with metastatic TNBC (mTNBC) is about one year from their metastatic diagnosis (Kassam 2009).

Although there is no evidence from randomised trials comparing chemotherapy with observation (i.e. no chemotherapy) in women with metastatic breast cancer, it is widely accepted that women with metastatic disease should receive some form of systemic therapy at some time during the course of their metastatic disease. Chemotherapy is considered by many to be the appropriate first treatment option for women with multiple sites of recurrence or where visceral disease is not easily treated by local modalities (Hayes 1995; Beslija 2009). Chemotherapy is also considered to be useful in women whose cancer is hormone refractory or is expected to be hormone resistant (Hortobagyi 1996).

Description of the intervention

Platinum compound, an alkylating agent, has been known to be active in metastatic breast cancer since clinical trials in the 1970s. However, it is more toxic and difficult to administer than other chemotherapy agents. The three most widely used platinum agents for treating breast cancer are cisplatin, carboplatin (both divalent complexes) and oxaliplatin (a tetravalent complex) (Sikov 2015). Cisplatin and carboplatin have demonstrated benefits in treating a number of cancer types including lung, testicular, head and neck, bladder and ovarian cancers. Oxaliplatin is often used to treat cisplatin‐ and carboplatin‐resistant tumours because it is commonly believed that cross‐resistance between oxaliplatin and cisplatin or carboplatin is incomplete (Mani 2002). More recent evidence suggests that the benefits of oxaliplatin may be due to its low toxicity and ability to be combined with other drugs rather than incomplete cross‐resistance with other platinum agents (Stordal 2007).

The use of oxaliplatin for treating breast cancers is much less common than the use of cisplatin or carboplatin, both in normal clinical practice and as an intervention in clinical trials (Sikov 2015). Cisplatin and carboplatin have been used and studied extensively as first‐line metastatic therapy in combination with other older pharmacological agents including 5‐fluorouracil and etoposide, and more recently with doxorubicin, epirubicin, vinorelbine, paclitaxel, docetaxel, cyclophosphamide, methotrexate and gemcitabine. The potential benefits of cisplatin or carboplatin as monotherapy for metastatic breast cancer, rather than as combination therapy, are rarely studied in clinical trials.

Although platinum agents have been shown to be efficacious in the treatment of a number of cancer types, their use is often associated with a variety of side effects. The known side effects of platinum agents include nausea, vomiting, myelosuppression (thrombocytopaenia, leukopaenia, neutropaenia and anaemia), peripheral neuropathy (symptoms include tingling in fingers and toes), nephrotoxicity, ototoxicities (hearing loss and tinnitus), hypomagnesaemia and anaphylaxis. Carboplatin is reported to be more tolerable than cisplatin with less nausea and vomiting, nephrotoxicity, ototoxicity and neurotoxicity, but worse myelosuppression, especially thrombocytopaenia (Sikov 2015).

How the intervention might work

The exact mechanism of action of platinum agents is not known but deoxyribonucleic acid (DNA) adducts are formed (Sikov 2015). These complexes are believed to inhibit DNA synthesis, replication and transcription by forming interstrand and intrastrand cross‐linking of DNA molecules. Interstrand cross‐links that remain intact can produce cell death, and it is this cytotoxic effect, when successful, that forms the mechanistic basis of action for cancer cell death by platinum agents (Noll 2006). For TNBC, it has been additionally hypothesised that a dysfunctional BRCA1 pathway in some TNBCs may make them more sensitive to platinum agents that selectively target cells deficient in homologous recombination DNA repair (Foulkes 2010).

Why it is important to do this review

In a previous Cochrane Review (Egger 2017), we found high‐quality evidence of little or no survival benefit from platinum‐based regimens for women with metastatic breast cancer unselected for triple‐negative disease. In that review we concluded that in relation to platinum agents, "... it is difficult to justify their use over commonly‐available less toxic active agents as first‐line treatment for metastatic patients without mTNBC." We determined that this conclusion was unlikely to change with the inclusion of additional studies. Hence, we are no longer updating the previous review in regards to women with metastatic breast cancer, unselected for triple‐negative disease.

However, in the previous review's subgroup analyses, we found preliminary low‐quality evidence of a survival benefit from platinum‐based regimens for women with mTNBC. This finding led us to conclude that "... although the evidence may be premature to recommend widespread use of platinum‐based regimens for mTNBC patients, some women and clinicians may consider platinum‐based regimens worth trying." We determined that there is a reasonable likelihood that this conclusion could change with the inclusion of additional studies; therefore we believe it is important to conduct a new review that updates the evidence from the mTNBC subgroup analyses in our previous Cochrane Review (Egger 2017).

Objectives

To assess the effects of platinum‐containing chemotherapy regimens with regimens not containing platinum in the management of women with mTNBC.

Additional objectives of this review were to investigate whether or not women in selected subgroups of studies benefited more or less from platinum‐based chemotherapy. Subgroups analyses were pre‐specified in the protocol of our previous review (Egger 2017), conducted in the original version of the review (Carrick 2004) or added in response to new hypotheses and the availability of new subgroups.

Methods

Criteria for considering studies for this review

Types of studies

Properly randomised controlled clinical trials (i.e. where the trial report asserts that the trial was randomised and there was no evidence to suggest otherwise) were eligible for inclusion. Because individual trials may compare one or more platinum‐based regimens to one or more non‐platinum‐based regimens, there were more 'treatment‐comparisons' (i.e. platinum regimen versus non‐platinum regimen comparisons) than studies in this review.

Types of participants

Participants are women with mTNBC, whether newly diagnosed or recurrent, who may have been purposely selected for mTNBC, or inadvertently selected as a subgroup. Treatment‐comparisons that included groups of women with loco‐regionally recurrent disease or women with non‐TNBC were only eligible for inclusion if it was possible to distinguish between these groups (i.e. where data were reported separately) or if the proportion of women in each group represented at least 80% of the total group. There were no age restrictions.

In the protocol for the previous review (Egger 2017), it was proposed that studies would be included if the women randomised to receive chemotherapy were to receive it as first‐line treatment (i.e. if no previous chemotherapy were given except as adjuvant therapy). As few studies assessing first‐line treatment were identified for inclusion in the original version of the previous review, those meeting the remaining eligibility criteria but which involved participants who were not first‐line naive were included. This modification of the inclusion criteria was maintained for this review, with subgroup analysis by treatment being performed (treatment‐comparisons with first‐line therapy for > 80% of participants versus other treatment lines).

Types of interventions

Interventions were any chemotherapy regimen containing a platinum agent (see Table 1 and Table 2). Comparators were any chemotherapy regimen without a platinum agent. In the protocol for the previous review (Egger 2017), endocrine therapy could also have been given to participants if it had been planned to be given to both treatment groups. However, endocrine therapy is unlikely to be relevant to women with TNBC.
Studies may or may not have specified recommended treatment upon disease progression or initial treatment failure, or both. This recommended treatment may have included cross‐over to the alternative treatment arm of the treatment‐comparison.

Table 1. Common platinum agents

Generic name	Other names
Carboplatin	Blastocarb, Carboplat, Carboplatin Hexal, Carboplatino, Carbosin, Carbosol, Carbotec, CBDCA, Displata, Ercar, Nealorin, Novoplatinum, Paraplat, Paraplatin AQ, Paraplatin, Paraplatine, Platinwas, Ribocarbo
Cisplatin	Abiplatin, Blastolem, Briplatin,CACP, CDDP, cis‐DDP, cis‐diamminedichloridoplatinum, cis‐diamminedichloro platinum (II), cis‐diamminedichloroplatinum, Cis‐dichloroammine Platinum (II), Cismaplat, Cisplatina, cis‐platinous diamine dichloride, cis‐platinum II diamine dichloride, cis‐platinum II, cis‐platinum, Cisplatyl, Citoplatino, Citosin, CPDD, Cysplatyna, DDP, DDP, Lederplatin, Metaplatin, Neoplatin, PDD, Peyrone's Chloride, Peyrone's Salt, Placis, Platamine, Platiblastin, Platiblastin‐S, Platinex, Platinol‐ AQ, Platinol, Platinol‐AQ VHA Plus, Platinol‐AQ, Platinoxan, platinum diamminodichloride, Platiran, Platistin, Platosin
Oxaliplatin	Ai Heng, Aiheng, diaminocyclohexane oxalatoplatinum, oxalatoplatin, oxalatoplatinum, oxaliplatine, Eloxatin, Dacotin, Dacplat, Eloxatine, 1‐OHP, L‐OHP, oxaliplatin medac

Table 2. Chemotherapeutic Agents (adapted from Table 1.1 in The Chemotherapy Source Book)

Type of Agent	Action	Includes
Agents that damage the DNA template	by alkylation: nitrogen mustards	cyclophosphamide, melphalan, ifosfamide, chlorambucil
	by alkylation: nitrosureas	carmustine (BCNU), lomustine (CCNU)
	by alkylation: other agents	thiotepa, mitomycin C
	by platinum coordination cross‐linking	cisplatin, carboplatin
	antibiotics	doxorubicin, daunorubicin, mitoxantrone, idarubicin, epirubicin, amsacrine
	podophyllotoxins	etoposide, teniposide
	by intercalation	dactinomycin, mithramycin
	by uncertain mechanisms	bleomycin
Spindle poisons	vinca alkaloids	vincristine, vinblastine, vendesine, vinorelbine
	taxanes	taxol, taxotere
Antimetabolites	thymidylate synthase	5‐fluorouracil
	dihydrofolate reductase	methotrexate

Types of outcome measures

Primary outcomes

Overall survival (OS)
Progression‐free survival/time to progression (PFS/TTP)

Secondary outcomes

Time to treatment failure (TTF)
Objective tumour response rate (OTRR)
Toxicity rates (multiple condition‐specific outcomes)
Quality of life (QoL) measures (multiple outcomes)

The definitions of some outcomes varied slightly across studies included in this review. Outcomes were commonly defined as the following.

OS: time elapsed between randomisation (or study enrolment or treatment initiation) to date of death from any cause.
Progression‐free survival (PFS): time elapsed between randomisation (or study enrolment or treatment initiation) and event, with event defined as disease progression or death from any cause.
Time to progression (TTP): time elapsed between randomisation (or study enrolment or treatment initiation) and event, with event defined as disease progression (which sometimes included cause‐specific death from the study disease).
TTF: time elapsed between randomisation (or study enrolment or treatment initiation) to treatment discontinuation for any reason, including disease progression, treatment toxicity, participant preference, or death.
OTRR: the proportion of participants who experienced a complete or partial tumour response (versus stable disease or no response).
Toxicity rates (multiple condition‐specific outcomes): the proportions of participants who experienced a grade 3 or 4 adverse event of nausea and vomiting, nephrotoxicity, anaemia, hair loss and leukopaenia, based on WHO criteria or individual protocol‐based definitions. We also investigated treatment‐related death which, for the purpose of this review, was defined as death due to the toxicity of the drug and not to disease progression or other cause. If an individual trial did not include their definition of a treatment‐related death but used the terms "toxic death" or "lethal toxicity," then these deaths were counted as treatment‐related deaths. Lastly, in response to a reviewer suggestion, we also examined treatment discontinuation due to adverse events.
QoL (generally measured using validated instruments for various QoL domains, but no studies in this review reported QoL results for mTNBC patients).

For the purposes of this review, we analysed PFS and TTP as the same outcome (referred to as PFS/TTP), with preference given to PFS for studies reporting both PFS and TTP data.

Search methods for identification of studies

Electronic searches

For this review, we searched the following databases and registries on the 27 September 2019.

The Cochrane Breast Cancer Specialised Register maintained by the Cochrane Breast Cancer Group (searched 2015 onwards). Details of the search strategies used by the Cochrane Breast Cancer Group for the identification of studies and the procedure used to code references are outlined in their module (www.mrw.interscience.wiley.com/cochrane/clabout/articles/BREASTCA/frame.html). Trials coded with the key words 'advanced,' 'Cisplatin,' 'cisplatinum,' 'carboplatin,' 'carboplatinum,' 'platin,' 'platinum,' 'platinum diamminodichloride,' 'cis‐diamminedichloroplatinum,' 'cis‐dichlorodiammineplatinum,' 'biocisplatinum,' 'dichlorodiammineplatinum,' 'nsc‐119875,' 'platidiam,' 'platino,' 'Platinol,' 'cis‐diamminedichloroplatinum,' 'cis‐platinum,' 'cis‐diammine (cyclobutanedicarboxylato) platinum,' 'cbdca,' 'jm‐8,' 'nsc‐241240,' 'paraplatin' were extracted for consideration.
Cochrane Central Register of Controlled Trials (CENTRAL; 2019, Issue 9) in the Cochrane Library. See Appendix 1.
MEDLINE (via OvidSP; from 2015 to 27 September 2019). See Appendix 2.
Embase (Via OvidSP; from 2015 to 27 September 2019). See Appendix 3.
The WHO International Clinical Trials Registry Platform (ICTRP) search portal (http://apps.who.int/trialsearch/Default.aspx) for all prospectively registered and ongoing trials. See Appendix 4.
ClinicalTrials.gov (http://clinicaltrials.gov/ct2/home). See Appendix 5.

We applied no restrictions based on language.

Searching other resources

We obtained relevant studies published prior to 2015 and their extracted results from the mTNBC subgroup analysis in our previous Cochrane Review (Egger 2017). We also searched for potentially relevant studies from previous trial reports, systematic reviews, and meta‐analyses.

Data collection and analysis

Selection of studies

Two review authors independently applied the selection criteria (including the quality of randomisation) to each reference identified by the search strategy while masked to the study results. Any discrepancies regarding eligibility or quality were resolved by consensus or adjudication from a third review author. Studies that may appear to have met the eligibility criteria, but which were deemed ineligible, are listed in the Characteristics of excluded studies table.

Data extraction and management

Data on the relevant outcomes were extracted by at least two review authors, with discrepancies resolved by consensus or adjudication from another review author. Data were also extracted on information relating to outcome definitions, study accrual, randomisation methods, baseline characteristics of participants (e.g. age; first‐line or second‐line treatment; prior anthracyclines or anthracycline‐naive), chemotherapy regimens (number of cycles and duration), follow‐up time and analytical methods used. Where available, multiple publications on the same study were obtained and the most complete report was assigned as the primary reference. In instances where a more recent publication was used in this review for a study that was included in our previous review (Egger 2017), the year of the reference ID was also updated. We entered data into the Cochrane Review Manager 5 (RevMan 2014) software, and we used this software for most statistical analyses.

Assessment of risk of bias in included studies

We assessed potential sources of bias for all included studies, using the first version of Cochrane's 'Risk of bias' assessment tool (Higgins 2011). At least two review authors independently evaluated the risk of bias for each treatment‐comparison and resolved discrepancies by consensus or adjudication from an additional reviewer. We sought clarification from authors if the published data provided inadequate information for the review. We assessed the 'Risk of bias' domains of random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and 'other bias.' For each included study, we assigned ratings of 'high,' 'low,' or 'unclear' risk of bias for each 'Risk of bias' domain, following criteria outlined in the 'Risk of bias' assessment tool (Higgins 2011).

Open‐label studies are common in phase III oncology trials because it is often difficult to conceal treatments from participants, care‐providers and outcome assessors (due to differences in toxicities and treatment schedules of various treatments, for example). However, because a lack of blinding can affect risk of bias in different ways for different outcomes, we assessed blinding of outcome assessment by dividing outcomes into two outcome classes: 1) OS and 2) outcomes other than OS and QoL. We made this division because, unlike other outcomes, assessment of OS is unlikely to be affected by non‐blinding.

We also divided the 'incomplete outcome data' risk of bias domain into two outcome classes: 1) time‐to‐event outcomes and 2) binary (i.e. dichotomous) outcomes. For time‐to‐event outcomes, we deemed risk of bias to be low, unclear, and high risk if time‐to‐event analysis was intention‐to‐treat (ITT), modified intention‐to‐treat (mITT) or per‐protocol, respectively. For the binary outcomes (OTRRs and toxicity rates), risk of bias was deemed low, unclear, and high risk if the highest percentage of randomised participants excluded from effect estimation was less than 10%, between 10% and 15%, or more than 15%, respectively.

For 'Risk of bias' domains that we divided into outcome classes, we made assessments for all studies known to be measuring the outcomes, regardless of results being reported in sufficient detail to be included in meta‐analysis or reported at all (e.g. a study might specify OS as an outcome in the study protocol but not report any results).

Measures of treatment effect

We analysed OS, PFS/TTP and TTF as time‐to‐event outcomes, for which the hazard ratio (HR) is the most appropriate measure of treatment effect. If reported, the HR and associated variance were extracted directly from the trial publication(s), and these were used to calculate observed (O) minus expected (E) numbers of events and logrank variance (V) for each treatment‐comparison using the methods described by Tierney 2007 or Parmar 1998. If not reported, we obtained O minus E and V indirectly from other available summary statistics or from data extracted from published Kaplan‐Meier curves using the methods described by Tierney 2007 or Parmar 1998. For studies that did not report the relevant effect estimates and required curve extraction, the numbers at risk were based on reported minimum and maximum follow‐up times. If these were not reported, minimum follow‐up was estimated as the time taken to complete treatment, and maximum follow‐up was estimated using the last event reported in the relevant time‐to‐event curve. These follow‐up estimates were recorded in the Characteristics of included studies table under 'Notes.' For the purposes of data extraction, we gave preference to time‐to‐event effect estimates derived from ITT analysis, followed by mITT analysis, then per‐protocol analysis.

We obtained pooled HRs and 95% CIs from the O minus E and V statistics for each treatment‐comparison, using the fixed‐effect model (Yusuf 1985). The pooled HR represented the instantaneous risk of an event (such as death, disease progression or treatment failure) for women receiving platinum, divided by the corresponding risk for those not receiving platinum. HRs less than 1.00 favoured the platinum‐containing regimens and values greater than 1.00 favoured non‐platinum regimens.

We analysed toxicity rates and OTRRs as proportions using the RR as the measure of treatment effect. OTRRs were most often calculated by trialists using only participants that were assessable for tumour response. These 'assessable participants' were generally defined as participants whose tumour response could be assessed according to prespecified criteria such as RECIST (Eisenhauer 2009); this definition was sometimes extended to additionally exclude participants who had not received a specified minimum dose of chemotherapy. In this review, we calculated OTRRs using the numbers of assessable participants in the OTRR denominators, where available, and randomised participants in the OTRR denominators where assessable participants were not available. Toxicity rates were most often calculated by trialists using a 'safety population' of participants who received a specified minimum dose of chemotherapy. We calculated toxicity rates for each study using the population used by that study.

We obtained pooled RRs and 95% CIs through Mantel‐Haenszel fixed‐effect analysis. The pooled RR represented the cumulative risk of an event for participants receiving platinum divided by the corresponding risk for those not receiving platinum. RRs greater than 1.00 favoured platinum‐containing regimens and values less than 1.00 favoured non‐platinum regimens.

QoL is generally reported as a continuous outcome. Hence, if sufficient QoL data become available for meta‐analysis in future review updates, the effect measure would most likely be the mean difference (MD) or standardized mean difference (SMD), depending on whether the same or different validated questionnaires (respectively) were employed. The direction of QoL scales will be standardized across individual studies such MD and SMD values greater than zero will favour (i.e. better QoL) platinum‐containing regimens while values less than zero will favour non‐platinum regimens. To help with interpretation, we will re‐express SMDs in the original units of one of the QoL instruments.

Unit of analysis issues

Treatment‐comparisons were the unit of analysis in this review and corresponded to pairwise comparisons of platinum and non‐platinum regimens. Individual studies assessing more than one platinum‐based regimen or more than one non‐platinum regimen (or both) contributed more than one treatment‐comparison to the review. Consequently, there were more treatment‐comparisons in this review than there were studies.

One study contained two non‐platinum regimen (control) groups for comparison against a single platinum‐based regimen (intervention) group. We took this into account when we calculated treatment effect statistics by splitting the study into two treatment‐comparisons (Stemmler 2011 A and Stemmler 2011 B) and halving the number of participants in the intervention group. For odd‐numbered group sizes, the additional participant was arbitrarily distributed to the treatment‐comparison with the label ending with 'A'. Two studies contained two platinum‐based regimen (intervention) groups for comparison against a single non‐platinum (control) group. These studies was split into two treatment‐comparisons (Han 2018 A and Han 2018 B; Yardley 2018 A and Yardley 2018 B) with treatment effect statistics calculated by halving the number of participants in the control group (with additional participants again arbitrarily distributed to treatment‐comparisons with label ending with 'A'). These methods for correcting for multiple intervention and/or control groups were suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2019).

Dealing with missing data

We made attempts to contact a number of trial investigators for additional information. One trialist (Icli 2005) provided additional results relating to their mTNBC subgroup.

Assessment of heterogeneity

Heterogeneity (variation) between trial results was assessed using the Chi² test statistic and the I² statistic. The Chi² test statistic assesses the amount of variation in a set of trials. Small P values for the Chi² test statistic suggest that there is more heterogeneity present than would be expected by chance. Chi² is not a particularly sensitive test: a cut‐off of P value less than 0.10 is often used to indicate significance, but lack of statistical significance does not mean there is no heterogeneity. I² is the proportion of variation that is due to heterogeneity rather than chance. In conjunction with the Chi² test, we used the I² statistic to assess heterogeneity using the rule of thumb guide outlined in the Cochrane Handbook (Higgins 2019) (i.e. I² between 0% to 40% might not be important; between 30% to 60% may represent moderate heterogeneity; between 50% to 90% may represent substantial heterogeneity; and between 75% to 100% considerable heterogeneity).

Assessment of reporting biases

In addition to assessing each treatment‐comparison individually for selective outcome reporting, using the first version of Cochrane's 'Risk of bias' tool (see Assessment of risk of bias in included studies above), we assessed publication bias and/or small‐study effects for the outcomes OS, PFS/TTP and OTRR by visual inspection of funnel plot asymmetry. We used Egger's statistical test to formally assess the degree of asymmetry (Egger 1997).

Data synthesis

For time‐to‐event outcomes, we used RevMan 5 (RevMan 2014) to estimate pooled HRs and 95% CIs, using fixed‐effect models of the derived or reported observed (O) and expected (E) number of events, and the variance of the log‐rank statistic (V) for each trial. For binary outcomes, we used RevMan 5 (RevMan 2014) to estimate pooled RRs and 95% CIs, using the fixed‐effect Mantel‐Haenszel method.

Subgroup analysis and investigation of heterogeneity

We performed subgroup analyses to determine whether the results differed by:

type of regimen comparison: (a) regimen A + platinum versus regimen A, (b) regimen A + platinum versus regimen B, (c) single agent platinum versus regimen C; (note that we allowed 'regimen A' to differ in dosage by small amounts between intervention and control arms);
type of platinum agent in platinum arm: (a) cisplatin, (b) carboplatin, (c) oxaliplatin;
first‐line therapy: (a) first‐line therapy for > 80% of participants, (b) second‐ or third‐line therapy for ≥ 20% of participants;
taxane in regimens: (a) no taxane in platinum or non‐platinum regimens, (b) platinum + taxane versus non‐platinum + taxane regimens, (c) platinum + non‐taxane versus non‐platinum + taxane regimens, (d) platinum + taxane versus non‐platinum + non‐taxane regimens;
BRCA1/2 mutation status: (a) germline BRCA1/2 mutation, (b) germline BRCA1/2 wild‐type; and
homologous recombination deficiency status: (a) homologous recombination deficient, (b) not homologous recombination deficient.

We assessed possible subgroup differences using Chi² tests.

Of the above six subgroup analyses:

subgroup analysis 1 was the only a priori subgroup analysis pre‐specified in the protocol of our previous review (Egger 2017); all other subgroup analyses were post hoc;
subgroup analyses 2 to 4 were conducted in the original version (Carrick 2004) of our previous review (Egger 2017), and in our previous review;
subgroup analysis 5 was added to the current review because three of the included trials (Han 2018 A/Han 2018 B Tutt 2018, Zhang 2018) hypothesised that BRCA1/2 positive breast cancers may be sensitive to chemotherapy regimens containing platinum; and
subgroup analysis 6 was added to the current review in response to hypotheses that somatic changes in tumours, similar to the effect of a germline BRCA mutation, could be predicted by HRD.

In the original version (Carrick 2004) of our previous review (Egger 2017), and in our previous review, we assessed trastuzumab in regimens as a subgroup analysis. This analysis, however, was not included in the current review because trastuzumab is not a relevant treatment for women with TNBC.

in order to reduce the number of forest plots in this review, toxicity rates were only shown overall and by subgroup analysis 2 ('type of platinum agent'). In general, however, few intervention‐comparisons could be included in meta‐analyses for adverse events. Better evidence of the toxicity of platinum regimens compared to non‐platinum regimens is found in our previous review's analysis of adverse events in women with metastatic breast cancer (Egger 2017), regardless of mTNBC status.

We had also intended to do a subgroup analysis looking at anthracycline in regimens. But because no studies had anthracycline in their platinum or non‐platinum regimens, the anthracycline subgroup meta‐analysis was not performed in this current version the review.

Sensitivity analysis

We performed several sensitivity analyses. First, we performed sensitivity analyses to assess whether the absence of OS data from some studies included in this review may have affected the OS result. In these analyses, we subgrouped the pooled effect estimates for OTRR and PFS/TTP according to whether treatment‐comparisons were included in OS meta‐analysis. Second, we stratified PFS/TTP estimates according to whether the outcome was PFS or TTP. For these analyses, we classified estimates as PFS if the event of interest was defined as disease progression or death from any cause. We classified estimates as TTP if the event of interest was defined as disease progression, which may also include cause‐specific death from breast cancer. In instances where the event of interest was ambiguously defined or not defined at all, we relied on the authors label of the outcome for classifying as PFS or TTP. Third, to assess the sensitivity of our primary results to our choice of analytical method, we repeated the main analyses (Analysis 1.1, Analysis 1.2 and Analysis 1.3) but using random‐effects rather than fixed‐effect methods. Fourth, it is plausible that trials not specifically assessing mTNBC patients might be more inclined to publish statistically significant mTNBC results from a subgroup analysis that was not pre‐specified in the trial protocol or trial registration. Consequently, we performed sensitivity analysis which subgrouped trials into those designed to specifically assess mTNBC patients and those where mTNBC patients were part of a post‐hoc subgroup analysis. Fifth, because Carey 2012 used a drug which is now widely considered to be ineffective in the treatment of breast cancer, the benefits of platinum chemotherapy are potentially exaggerated in the Carey 2012 trial. Therefore, we repeated the main analyses after exclusion of Carey 2012.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE system (Guyatt 2011) to rate the quality of evidence relating to the estimated treatment effects on OS, PFS/TTP and OTRR, as well as on rates of treatment‐related death, nausea/vomiting, anaemia, hair loss, leukopaenia and treatment discontinuation due to adverse events. GRADE criteria for assessing quality of evidence include study design, risk of bias, inconsistency, indirectness, imprecision, suspected publication bias and other considerations. Assessments of these criteria and corresponding justifications are provided in three 'Summary of findings' tables, largely created using GRADEproGDT (GradeproGDT). We performed GRADE assessments separately for selected subgroups related to inconsistency (i.e. heterogeneity) among effect estimates.

Results

Description of studies

Results of the search

We reviewed 1199 unique records identified by the 2019 database searches (Figure 1). Of these, we excluded 1177 based on information in the title or abstract. We considered 11 records from trial registries or protocol publications to be potentially relevant. These were ongoing studies that have not yet published results (see Characteristics of ongoing studies). For the remaining 11 records, we retrieved full‐text articles or abstracts for further examination. We excluded five of the 11 articles or abstracts because they were review articles, but we examined their bibliographies to search for additional relevant studies. We excluded one other full‐text article for reasons outlined in the Characteristics of excluded studies table. We re‐assessed studies included in our previous review (Egger 2017) for inclusion in the current review.

Figure 1

Review 2020: study flow diagram.

Included studies

We included 10 studies with 13 treatment‐comparisons in this review.

Of these 13 treatment‐comparisons, eight were identified from our previous review (Egger 2017): three with the same mTNBC‐specific results reported in our previous review (Bhattacharyya 2009; Carey 2012; Fan 2012) and five with new or updated mTNBC‐specific results (Icli 2005; Stemmler 2011 A; Stemmler 2011 B; Tutt 2018; Zhang 2018). The other five treatment‐comparisons were identified by our September 2019 search (Han 2018 A; Han 2018 B; Mustafa 2019; Yardley 2018 A; Yardley 2018 B).

Of the 13 treatment‐comparisons included in this review (Table 3):

Table 3. Number of treatment‐comparisons by subgroup and three outcomes

		Outcome
Subgroup	Treatment‐ comparisons N	Overall survival n (% of N)	Progression ‐free survival/time to progression n (% of N)	Objective tumour response rate n (% of N)
Overall:	13	6 (46%)	8 (62%)	10 (77%)
Type of regimen comparison:
Regimen A + platinum agent vs regimen A	2	1 (50%)	1 (50%)	2 (100%)
Regimen A + platinum agent vs regimen B	9	4 (44%)	5 (56%)	7 (78%)
Single agent platinum vs regimen C	2	1 (50%)	2 (100%)	1 (50%)
Type of platinum agent in platinum arm:
Cisplatin in platinum arm	7	2 (29%)	2 (29%)	6 (86%)
Carboplatin in platinum arm	6	4 (67%)	6 (100%)	4 (67%)
First‐line therapy:
First‐line therapy for > 80% of patients	6	5 (83%)	5 (83%)	6 (100%)
Second‐ or third‐line therapy for >=20% of patients	7	1 (14%)	3 (43%)	4 (57%)
Anthracycline in regimens:
No anthracycline in platinum or non‐platinum regimens	13	6 (46%)	8 (62%)	10 (77%)
Taxane in regimens:
No taxane in platinum or non‐platinum regimens	4	1 (25%)	1 (25%)	4 (100%)
Platinum + taxane vs non‐platinum + taxane regimens	2	2 (100%)	2 (100%)	2 (100%)
Platinum + non‐taxane vs non‐platinum + taxane regimens	5	3 (60%)	3 (60%)	4 (80%)
Platinum + taxane vs non‐platinum + non‐taxane regimens	2	(0%)	2 (100%)	(0%)
BRCA1/2 subtype:
Germline BRCA1/2 mutation #	4	1 (25%)	4 (100%)	2 (50%)
Germline BRCA1/2 wild‐type #	2	1 (50%)	2 (100%)	2 (100%)
Homologous recombination deficiency status:
Homologous recombination deficient #	2	1 (50%)	2 (100%)	2 (100%)
Not homologous recombination deficient #	2	1 (50%)	2 (100%)	2 (100%)

^Numbers for each outcome are the number of treatment‐comparison with sufficient data to be included in meta‐analysis for that outcome.

# BRCA1/2 subtype and homologous recombination deficiency status were within‐study subgroupings for Tutt 2018 and Zhang 2018; hence Tutt 2018 and Zhang 2018 both contributed to both BRCA1/2 subroups and both homologous recombination deficiency status subgroups.

two (15%) compared 'regimen A + platinum versus regimen A,' nine (69%) compared 'regimen A + platinum versus regimen B' and 2 (15%) compared single agent platinum versus regimen C' (Table 3);
seven (54%) used cisplatin and 6 (46%) used carboplatin as the platinum agent in the intervention arm (Table 3 and Table 4);
six (46%) had more than 80% of participants receiving first‐line therapy;
all 13 (100%) had no anthracycline in the platinum or non‐platinum regimens (consequently, the anthracycline subgroup meta‐analysis was not performed);
four (31%) had no taxane in the platinum or non‐platinum regimens; two (15%) had a taxane in both regimens, five (38%) had a taxane in the non‐platinum regimen only and two (15%) had a taxane in the platinum regimen only;
four (31%) had germline BRCA1/2 mutation subgroup results and two (13%) had BRCA1/2 wild‐type subgroup results; and
two (15%) were homologous recombination deficient and two (15%) were not homologous recombination deficient.

Table 4. Summary of regimens included in the analysis

Trials ID	Arm 1(platinum‐containing)	Arm 2(control)	First‐line therapy for > 80% of participants	Majority participants anthracycline‐naive
Regimen A + platinum vs regimen A
Bhattacharyya 2009	(Endoxan + with 'cisplatinum')	(Endoxan)	N	N
Carey 2012	C + Cb (Cetuximab + carboplatin)	C (Cetuximab with carboplatin added after progression)	N	N
Regimen A + platinum vs regimen B
Fan 2012	TP (docetaxel + cisplatin)	TX (docetaxel + capecitabine)	Y	N
Mustafa 2019	(cisplatin + gemcitabine)	(paclitaxel + gemcitabine)	Y	N
Han 2018 A	PCP (placebo + carboplatin/paclitaxel)	VT (veliparib + temozolomide)	N	Unknown
Han 2018 B	VCP (eliparib + carboplatin/paclitaxel)	VT (veliparib + temozolomide)	N	Unknown
Stemmler 2011 A	GemCis (gemcitabine + cisplatin)	GemVin (gemcitabine + vinorelbine)	N	N
Stemmler 2011 B	GemCis (gemcitabine + cisplatin)	GemCap (gemcitabine + capecitabine)	N	N
Yardley 2018 A	nab‐P/C (nab‐paclitaxel + carboplatin)	nab‐P/G (nab‐paclitaxel + gemcitabine)	Y	N
Yardley 2018 B	G/C (gemcitabine + carboplatin)	nab‐P/G (nab‐paclitaxel + gemcitabine)	Y	N
Zhang 2018	(cisplatin + gemcitabine)	(paclitaxel + gemcitabine)	Y	N
Icli 2005	Etop + Cis (etoposide + cisplatin)	P (paclitaxel)	N	N
Single agent platinum vs regimen C
Tutt 2018	C (carboplatin)	D (docetaxel)	Y	Y

Not all studies provided sufficient information on all outcomes for inclusion in meta‐analyses. Of the 13 treatment‐comparisons:

six (46%), eight (62%) and 10 (77%) had sufficient data to be included in the meta‐analyses of effect estimates for OS, PFS/TTP and OTRR, respectively (Table 3 and Figure 1); and
five (38%), three (23%), zero (0%), five (38%), two (15%), five (38%) and five (38%) had sufficient data to be included in the meta‐analyses of effect estimates for treatment‐related death, nausea/vomiting, nephrotoxicity, anaemia, hair loss, leukopaenia, and treatment discontinuation due to adverse events, respectively (Table 5 and Figure 1).

See: Summary of findings 1 Platinum compared to non‐platinum regimens for metastatic triple‐negative breast cancer: OS, PFS/TTP and OTRR; Summary of findings 2 Platinum‐containing regimens and toxicity profile

Table 5. Summary of outcomes for included trials

Trial ID	OS data useable for HR estimation for mTNBC patients¹	MedianOS time for mTNBC patients²	PFS/TTP data useable for HR estimation for mTNBC patients¹	Median PFS/TTP time for mTNBC patients²	Objective tumour response for mTNBC patients	Treatment‐related deaths for mTNBC patients	Grade III & IV Toxicity for mTNBC patients	Analysed³
Regimen A + platinum vs regimen A
Bhattacharyya 2009	NR	Y	NR	Y	Y	NR	NR	126
Carey 2012	Y	Y	Y	NR	Y	NR	Not useable for meta‐analysis	102
Regimen A + platinum vs regimen B
Fan 2012	Y	Y	Y	Y	Y	Y	Nausea/vomiting Anaemia Leukopaenia Treatment‐discontinuation	53
Mustafa 2019	NR	NR	NR	Y	Y	NR	Not useable for meta‐analysis	110
Han 2018 A	NR	NR	Y	NR	NR	NR	NR	60
Han 2018 B	NR	NR	Y	NR	NR	NR	NR	59
Stemmler 2011 A	NR	NR	NR	NR	Y	NR	NR	15
Stemmler 2011 B	NR	NR	NR	NR	Y	NR	NR	21
Yardley 2018 A	Y	Y	Y	Y	Y	Y	Anaemia Leukopaenia Treatment‐discontinuation	95
Yardley 2018 B	Y	Y	Y	Y	Y	Y	Anaemia Leukopaenia Treatment‐discontinuation	96
Zhang 2018	Y	NR	Y	Y	Y	Y	Nausea/vomiting Anaemia Hair loss Leukopaenia Treatment‐discontinuation	236
Icli 2005	NR	Y	NR	Y	N	Y	Nausea/vomiting Anaemia Leukopaenia	0
Single agent platinum vs regimen C
Tutt 2018	Y	Y	Y	Y	Y	Y	Nausea/vomiting Anaemia Hair loss Leukopaenia Treatment‐discontinuation	376

¹Sufficient data reported to estimate a HR for pooling as outlined by Parmar 1998 and Tierney 2007; this includes Kaplan‐Meier curve, HR and standard error/confidence interval or logrank statistics
²Trials that did not explicitly report median time were classified as NR here regardless of estimable median time from Kaplan‐Meier curve
³Analysed numbers represent the maximum numbers of participants in the treatment‐comparison that were included in a meta‐analysis of OS, PFS/TTP or OTRR (assessable participants).

DU: deaths unexplained
NR: not reported at all or not reported for mTNBC subgroup
OS: overall survival
PFS: progression‐free survival
TTP: time to progression
Y: year reported

No studies reported QoL or TTF results for women with mTNBC.

Excluded studies

Eight studies may have appeared to have met the eligibility criteria but were deemed ineligible for reasons given in the Characteristics of excluded studies table.

Risk of bias in included studies

Figure 2 shows a summary of the 'Risk of bias' judgements for each 'Risk of bias' domain of the included treatment‐comparisons. Reasons for each judgement are detailed for each treatment‐comparison in the Characteristics of included studies table. For each 'Risk of bias' domain, a summary of the general risk of bias for results of the included studies was as follows.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

All 10 studies, reporting 13 treatment‐comparisons, were described as randomised. The method of random sequence generation was described sufficiently to be judged at low risk of bias for this domain in six treatment‐comparisons (Bhattacharyya 2009; Carey 2012; Tutt 2018; Yardley 2018 A; Yardley 2018 B; Zhang 2018). The remaining seven treatment‐comparisons were judged to be at unclear risk of bias for random sequence generation, as the information available was insufficient to accurately assess this domain.

Five of the 13 treatment‐comparisons described central randomisation systems, and were thus judged to be at low risk of bias for treatment allocation concealment (Icli 2005; Tutt 2018; Yardley 2018 A; Yardley 2018 B; Zhang 2018). The remaining seven treatment‐comparisons did not adequately describe methods of concealment and were thus judged as having unclear risk of bias for this domain.

Blinding

Eleven treatment‐comparisons were described as "nonblinded," "not blinded," "single blind" or "open‐label" (Carey 2012; Fan 2012; Han 2018 A; Han 2018 B; Icli 2005; Stemmler 2011 A; Stemmler 2011 B; Tutt 2018; Yardley 2018 A; Yardley 2018 B; Zhang 2018). These 11 'unblinded' treatment‐comparisons were judged to be at high risk of 'performance bias' due to the lack of blinding of participants and personnel to the treatment being administered. The remaining two treatment‐comparisons were judged as at unclear risk of performance bias because of a lack of information needed to make a firm conclusion. It seemed highly likely, however, that these two treatment‐comparisons would have also been 'unblinded,' as open‐label studies are common in phase III oncology trials.

All 12 treatment‐comparisons known to have OS as a study outcome (including six not included in OS meta‐analyses; but excluding Mustafa 2019, which did not assess OS as an outcome) were judged to be at low risk of bias from a lack of blinding of outcome assessors, regardless of actual blinding. This is because death certification was unlikely to have been affected by any lack of blinding.

For outcomes other than OS and QoL, two treatment‐comparisons were judged to be at low risk of bias from a lack of blinding of outcome assessors due to these outcomes being measured or confirmed through formal assessments including imaging, biochemical tests and/or the involvement of an independent clinical or radiological review group (Carey 2012; Icli 2005). Three treatment‐comparisons were judged to be at high risk of bias from a lack of blinding of outcome assessors (Tutt 2018; Yardley 2018 A; Yardley 2018 B). The remaining eight treatment‐comparisons provided insufficient detail on outcome assessments and were thus classified as having an unclear risk of bias.

Incomplete outcome data

Five treatment‐comparisons excluded randomised participants who never started treatment or who were subsequently found to have been 'ineligible' from time‐to‐event analyses (mITT analyses) (Carey 2012; Han 2018 A; Han 2018 B; Icli 2005; Zhang 2018). These five treatment‐comparisons were judged to be at unclear risk of attrition bias for time‐to‐event outcomes. The remaining eight treatment‐comparisons were judged to be at low risk of attrition bias for time‐to‐event outcomes because all randomised participants were analysed in the groups to which they were randomised (ITT analysis).

Two treatment‐comparisons had more than 15% of participants not assessed or not assessable for at least one binary outcome, and were thus judged to be at high risk of attrition bias for binary outcomes (Stemmler 2011 A; Stemmler 2011 B). Five treatment‐comparisons had less than 10% of participants not assessed or not assessable for all binary outcomes, and were thus judged to be at low risk of attrition bias for binary outcomes (Fan 2012; Mustafa 2019; Yardley 2018 A; Yardley 2018 B; Zhang 2018). The remaining six treatment‐comparisons were judged to be at unclear risk of attrition bias for binary outcomes (10% to 15% of participants not assessed or not assessable for at least one binary outcome, or it was unclear what proportion were not assessed).

Selective reporting

The assessment of risk of bias from selective reporting included cross‐checking the outcomes for which there were published results against the stated outcomes reported in trial registers and published protocols. In our assessment of risk of bias from selective reporting, studies that began recruiting participants on or after July 1, 2005 were expected have a clinical registration or published protocol specifying the study outcomes, or we deemed them to be at high risk of bias from selective reporting. We chose July 1, 2005 as our early limit because the International Committee of Medical Journal Editors (ICMJE) made a seminal announcement in September 2004 that clinical trials that begin recruiting on or after July 1, 2005 would not be considered for publication unless they were included on a clinical trials registry (De Angelis 2005). Studies included in this review that began recruiting participants before July 1, 2005 and which did not have a trial registration or published protocol pre‐specifying study outcomes, were assumed to be at unclear risk of bias from selective reporting, unless additional evidence suggested otherwise.

Four treatment‐comparisons from three studies were judged to be at low risk of bias from the selective reporting of outcomes (Carey 2012; Yardley 2018 A; Yardley 2018 B; Zhang 2018). Each of these studies was included on a clinical trials registry and their prespecified outcomes either matched those in the trial reports or non‐matches were considered to be relatively minor. Eight treatment‐comparisons were judged to be at high risk of bias from the selective reporting of outcomes. Of these eight treatment‐comparisons: Bhattacharyya 2009 indicated in the abstract that toxicity was recorded, but did not report results; there was no trial registration or published protocol containing the study's prespecified outcomes). Fan 2012 and Mustafa 2019 did not have a trial registration or published protocol, despite recruitment beginning after July 1, 2005. Han 2018 A/Han 2018 B, Icli 2005 and Stemmler 2011 A/Stemmler 2011 B did not report all outcomes specified in their protocol for the mTNBC subgroup analysis. We judged the remaining treatment‐comparison (Tutt 2018) to be at unclear risk of bias from the selective reporting of outcomes because while the protocol‐specified outcomes of TTP and TTF were not reported, the similar outcome of PFS was reported.

Egger's tests for funnel plot asymmetry indicated some evidence consistent with the presence of publication bias or small‐study effects, or both, for PFS/TTP (P = 0.02; Figure 3) but not for OS (P = 0.08; Figure 4) or OTRR (P = 0.12; Figure 5).

Figure 3

Funnel plot for PFS/TTP (Progression‐free survival/time to progression). Assessing publication bias and/or small‐study effects. Plot includes all treatment‐comparisons with data for PFS/TTP that could be included in meta‐analysis. The plot suggests some level of asymmetry (Egger's test P value = 0.02).

Figure 4

Funnel plot for overall survival (OS). Assessing publication bias and/or small‐study effects. Plot includes all treatment‐comparisons with data for OS that could be included in meta‐analysis.. The plot does not show substantial asymmetry (Egger's test P value = 0.08)

Figure 5

Funnel plot for objective tumour response rate (OTRR). Assessing publication bias and/or small‐study effects. Plot includes all treatment‐comparisons with data for OTRR that could be included in meta‐analysis. The plot does not show asymmetry (Egger's test P value = 0.12).

Other potential sources of bias

Eight treatment‐comparisons were judged to be at unclear risk of 'other bias' (Carey 2012; Han 2018 A; Han 2018 B; Mustafa 2019; Tutt 2018; Yardley 2018 A; Yardley 2018 B; Zhang 2018) for various reasons outlined in the Characteristics of included studies table. The remaining five treatment‐comparisons were judged to be at low risk of 'other bias.'

Effects of interventions

Please refer to summary of findings Table 1

Overall survival

Twelve of the 13 included treatment‐comparisons assessed OS as an outcome; six provided sufficient OS data specific to mTNBC patients for pooling in meta‐analyses. From these six treatment‐comparisons, 958 of 972 randomised participants were analysed representing 99% of randomised participants in these treatment‐comparisons (and there were about 573 deaths). Pooled analysis indicated a 15% lower rate of death for women receiving platinum‐containing regimens compared to those receiving non‐platinum regimens (HR 0.85, 95% CI 0.73 to 1.00; P = 0.05; heterogeneity P = 0.41, I² = 1%; moderate‐quality evidence) (Analysis 1.1; Figure 6). Subgroup analyses of OS indicated no evidence of subgroup differences (P values ranged from P = 0.19 to P = 0.89; see Analysis 2.1; Analysis 3.1; Analysis 4.1; Analysis 5.1; Analysis 6.1; Analysis 7.1).

Figure 6

Forest plot of comparison: 1 Platinum vs non‐platinum regimens, outcome: 1.1 Overall survival.

PFS/TTP

All 13 included treatment‐comparisons assessed PFS or TTP, or both, as an outcome; eight provided sufficient mTNBC‐specific data for pooling in meta‐analyses of the composite outcome of PFS/TTP. From these eight treatment‐comparisons, 1077 out of 1092 (99%) randomised participants were analysed (with approximately 909 events). Pooled analysis indicated platinum‐containing regimens were associated with better PFS/TTP (HR 0.77, 95% CI 0.68 to 0.88; P < 0.0001), although the quality of the evidence was very low. This was due to marked evidence of heterogeneity (P < 0.0001; I² = 80%) (Analysis 1.2; Figure 7), the use of PFS/TTP is a surrogate endpoint and suspected publication bias (summary of findings Table 1).

Figure 7

Forest plot of comparison: 1 Platinum vs non‐platinum regimens, outcome: 1.2 Progression‐free survival/time to progression.

Evidence of subgroup differences in the pooled HRs of subgroups was found in four of the six subgroup analyses involving PFS/TTP:

comparing 'regimen A + platinum versus regimen A' (HR 0.56, 95% CI 0.41 to 0.77; n = 1); 'regimen A + platinum versus regimen B' (HR 0.68, 95% CI 0.55 to 0.82; heterogeneity P = 0.007, I² = 71%; n = 5); and 'single agent platinum versus regimen C' (HR 1.00, 95% CI 0.83 to 1.22; heterogeneity P = 0.004, I² = 88%; n = 2) (P = 0.002 for subgroup difference) (Analysis 2.2);
comparing first‐line therapy for > 80% of participants (HR 0.89, 95% CI 0.77 to 1.04; heterogeneity P = 0.002, I² = 77%; n = 5); and second‐ or third‐line therapy for ≥ 20% of participants (HR 0.50, 95% CI 0.38 to 0.64; heterogeneity P = 0.32, I² = 13%; n = 3) (P < 0.0001 for subgroup difference) (Analysis 4.2);
with no taxane in platinum or non‐platinum regimens (HR 0.56, 95% CI 0.41to 0.77; n = 1); a taxane in both the platinum and non‐platinum regimens (HR 0.46, 95% CI 0.30 to 0.70; heterogeneity P = 0.11, I² = 61%; n = 2); a taxane in the non‐platinum regimen only (HR 0.98, 95% CI 0.84 to 1.15; heterogeneity P = 0.17, I² = 43%; n = 3); and a taxane in the platinum regimen only (HR 0.37, 95% CI 0.23 to 0.59; heterogeneity P = 0.67, I² = 0%; n = 2) (P < 0.00001 for subgroup difference) (Analysis 5.2);
with women with germline BRCA 1/2 mutation (HR 0.43, 95% CI 0.30 to 0.62; heterogeneity P = 0.73, I² = 0%; n = 4); compared to women with germline BRCA 1/2 wild‐type (HR 1.14, 95% CI 0.93 to 1.40; heterogeneity P = 0.18, I² = 45%; n = 2). (P < 0.00001 for subgroup difference) (Analysis 6.2).

The two other subgroup analyses showed no evidence of subgroup differences (P values ranged from P = 0.14 to P = 0.75; see Analysis 3.2; Analysis 7.2).

TTF

None of the 13 included treatment‐comparisons assessed TTF as an outcome.

OTRR: assessable participants

All 13 included treatment‐comparisons assessed OTRR as an outcome; 10 provided sufficient OTRR data specific to mTNBC patients for pooling in meta‐analyses. From the 10 treatment‐comparisons, 1205 out of 1244 (97%) randomised participants were assessable for tumour response (and 529 had a complete or partial response). Women receiving platinum‐containing regimens had 40% better OTRR than women receiving non‐platinum regimens (RR 1.40, 95% CI 1.22 to 1.59, P < 0.00001), but the quality of the evidence was low because of evidence of heterogeneity (P = 0.01; I² = 58%) (Analysis 1.3; Figure 8) and because OTRR is a surrogate endpoint (summary of findings Table 1).

Figure 8

Forest plot of comparison: 1 Platinum vs non‐platinum regimens, outcome: 1.3 Objective tumour response rate (assessable participants).

Evidence of subgroup differences in the pooled RRs of subgroups were found in five of the six OTRR subgroup analyses:

comparing 'regimen A + platinum versus regimen A' (RR 2.14, 95% CI 1.42 to 3.23; heterogeneity P = 0.64, I² = 0%; n = 2); 'regimen A + platinum versus regimen B' (RR 1.51, 95% CI 1.29 to 1.77; heterogeneity P = 0.14, I² = 38%; n = 7); and 'single agent platinum versus regimen C' (RR 0.92, 95% CI 0.69 to 1.23; n = 1) (P = 0.003 for subgroup difference) (Analysis 2.3);
using cisplatin (RR 1.61, 95% CI 1.36 to 1.89; heterogeneity P = 0.05, I² = 54%; n = 6); and carboplatin (RR 1.16, 95% CI 0.93 to 1.44; heterogeneity P = 0.06, I² = 59%; n = 4) (P = 0.02 for subgroup difference) (Analysis 3.3);
comparing first‐line therapy for > 80% of participants (RR 1.30, 95% CI 1.13 to 1.50; heterogeneity P = 0.02, I² = 63%; n = 6); and second‐ or third‐line therapy for ≥ 20% of participants (RR 2.05, 95% CI 1.42 to 2.96; heterogeneity P = 0.25, I² = 27%; n = 4) (P = 0.02 for subgroup difference) (Analysis 4.3);
with no taxane in the platinum or non‐platinum regimens (RR 2.05, 95% CI 1.42 to 2.96; heterogeneity P = 0.25, I² = 27%; n = 4); a taxane in both platinum and non‐platinum regimens (RR 2.23, 95% CI 1.48 to 3.38; heterogeneity P = 0.11, I² = 61%; n = 2); a taxane in the non‐platinum regimen only (RR 1.18, 95% CI 1.02 to 1.38; heterogeneity P = 0.13, I² = 47%; n = 4) (P = 0.001 for subgroup difference) (Analysis 5.3);
with women with germline BRCA 1/2 mutation (RR 2.09, 95% CI 1.17 to 3.72; heterogeneity P = 0.89, I² = 0%; n = 2); compared to women with germline BRCA 1/2 wild‐type (RR 0.90, 95% CI 0.71 to 1.15; heterogeneity P = 0.05, I² = 74%; n = 2) (P = 0.008 for subgroup difference) (Analysis 6.3).

The other subgroup analyses showed no evidence of subgroup differences (P = 0.30 for Analysis 7.3).

Toxicity: safety populations

Please refer to summary of findings Table 2

Treatment‐related death

Five of the 13 included treatment‐comparisons reported treatment‐related death for mTNBC patients and provided sufficient data for extraction. Of these five treatment‐comparisons, two had non‐estimable RRs due to no treatment‐related deaths and thus did not contribute to the pooled estimates. For the three remaining treatment‐comparisons, 554 out of 567 (98%) randomised women were included in the safety populations, with six treatment‐related deaths. There was no evidence of a difference between platinum and non‐platinum regimens in terms of treatment‐related death but the quality of evidence was low because the confidence interval was very wide (RR 1.06, 95% CI 0.24 to 4.61; Analysis 3.4) (summary of findings Table 2). There was no evidence of heterogeneity (P = 0.69; I² = 0%).

It was not possible to perform subgroup analyses according to the type of platinum agent used, as the two cisplatin treatment‐comparisons had non‐estimable RRs, due to no treatment‐related deaths.

Nausea/vomiting

Three of the 13 included treatment‐comparisons reported grade 3 and 4 nausea/vomiting for mTNBC patients with sufficient data for extraction. Of these three treatment‐comparisons, 655 out of 669 (98%) randomised women were included in the safety populations with 31 cases of grade 3 or 4 nausea/vomiting. Risk of grade 3 or 4 nausea/vomiting was nearly five times higher among women receiving platinum‐containing regimens (RR 4.77, 95% CI 1.93 to 11.81; P = 0.0007) but the quality of evidence was deemed low because the confidence interval was very wide. There was no evidence of heterogeneity (P = 0.32, I² = 12%) (Analysis 3.5).

There was little evidence of difference in pooled RRs according to the type of platinum agent used (P = 0.15).

Nephrotoxicity

None of the 13 included treatment‐comparisons reported grade 3 and 4 nephrotoxicity for mTNBC patients.

Anaemia

Five of the 13 included treatment‐comparisons reported grade 3 and 4 anaemia in mTNBC patients and provided sufficient data for extraction. Of these five treatment‐comparisons, one had a non‐estimable RR due to no grade 3 and 4 anaemia and thus did not contribute to the pooled estimates. For the three remaining treatment‐comparisons, 790 out of 807 (98%) randomised women were included in the safety populations, with 86 cases of grade 3 and 4 anaemia. Risk of grade 3 or 4 anaemia was nearly four times higher among women receiving platinum‐containing regimens (RR 3.80, 95% CI 2.25 to 6.42; P < 0.00001) (Analysis 3.6), but the quality of evidence was deemed low because the confidence interval was wide and there was some evidence of heterogeneity (P = 0.04; I² = 65%).

Subgroup analysis showed weak evidence indicating that the increased risk of grade 3 or 4 anaemia for platinum recipients (compared to non‐platinum recipients) was worse for cisplatin recipients compared to carboplatin recipients (P = 0.06).

Hair loss

Two of the 13 included treatment‐comparisons assessed grade 3 and 4 hair loss for mTNBC patients and reported sufficient data for extraction. One of these two treatment‐comparisons had a non‐estimable RR due to there being no grade 3 or 4 cases. For the one remaining treatment‐comparison, 366 out of 376 (97%) randomised women were included in the safety population, with one case of grade 3 or 4 hair loss. There was no evidence that the risk of grade 3 or 4 hair loss was different for women receiving platinum‐containing regimens (RR 0.33, 95% CI 0.01 to 8.04) (Analysis 3.7), but the quality of evidence was low because the confidence interval was very wide.

Leukopaenia

Five of the 13 included treatment‐comparisons assessed grade 3 and 4 leukopaenia for mTNBC patients and reported sufficient data for extraction. One treatment‐comparison had a non‐estimable RR due to there being no grade 3 or 4 cases of leukopaenia. For the four remaining treatment‐comparisons, 790 out of 807 (98%) randomised women were included in the safety populations, with 131 cases of grade 3 or 4 leukopaenia. There was no evidence of a difference in risk of grade 3 or 4 leukopaenia between platinum and non‐platinum containing regimens (RR 1.09, 95% CI 0.84 to 1.42, P = 0.52). The quality of evidence was deemed moderate because the confidence interval was wide. There was no evidence of heterogeneity (P = 0.75, I² = 0%) (Analysis 3.8).

There was no evidence of differences in pooled RRs according to the type of platinum agent used (P = 0.44).

Treatment discontinuation due to adverse events

Five of the 13 included treatment‐comparisons assessed treatment discontinuation due to adverse events and reported sufficient data for extraction. One treatment‐comparison had a non‐estimable RR due to there being no treatment discontinuations. For the four remaining treatment‐comparisons, 790 out of 807 (98%) randomised women were included in the safety populations, with 89 treatment discontinuations. There was no evidence of a difference in risk of treatment discontinuations between platinum and non‐platinum containing regimens (RR 0.88, 95% CI 0.59 to 1.32, P = 0.55). The quality of evidence was deemed moderate because the confidence interval was wide. There was some evidence of heterogeneity (P = 0.07, I² = 57%) (Analysis 3.9).

There was no evidence of differences in pooled RRs according to the type of platinum agent used (P = 0.63).

QoL

None of the 13 included treatment‐comparisons reported QoL outcomes for mTNBC patients.

Sensitivity analyses

The benefit of platinum regimens over non‐platinum regimens in terms of PFS/TTP was more pronounced for treatment‐comparisons not included in the OS meta‐analysis (HR 0.37, 95% CI 0.23 to 0.59; n = 2) than for treatment‐comparisons included the OS meta‐analysis (HR 0.82, 95% CI 0.72 to 0.94; n = 6) (P = 0.001 for subgroup difference) (Analysis 8.1). Similarly, the benefit of platinum regimens over non‐platinum regimens in terms of OTRR was marginally more pronounced for treatment‐comparisons not included in the OS meta‐analysis (RR 1.70, 95% CI 1.33 to 2.18; n = 4) than for treatment‐comparisons included the OS meta‐analysis (RR 1.30, 95% CI 1.11 to 1.52; n = 6) (P = 0.07 for subgroup difference) (Analysis 8.2). Together these results provide some suggestion that the absence of some treatment‐comparisons (namely Bhattacharyya 2009; Han 2018 A; Han 2018 B; Mustafa 2019; Stemmler 2011 A; Stemmler 2011 B) from the OS meta‐analysis may have, if anything, lead to an underestimate of the benefit of platinum regimens in terms of OS.
Stratifying PFS/TTP estimates according to whether the outcome was PFS or TTP suggested that platinum chemotherapy was more beneficial in terms of TTP than PFS (P = 0.03), although this difference was based on only one treatment‐comparison in the TTP group (Analysis 9.1).
Repeating Analysis 1.1, Analysis 1.2 and Analysis 1.3 using random‐effects methods did not appreciably change the pooled estimates (Analysis 10.1; Analysis 10.2 and Analysis 10.3).
Of the 13 treatment‐comparisons included in this review, five were results from trials not specifically assessing mTNBC patients (i.e. mTNBC patients were part of a subgroup analysis; Han 2018 A; Han 2018 B; Icli 2005; Stemmler 2011 A; Stemmler 2011 B). Of these five treatment‐comparisons, four were included in meta‐analyses (Han 2018 A and Han 2018 B were included in meta‐analyses for PFS/TTP and Stemmler 2011 A and Stemmler 2011 B were included in meta‐analyses for OTRR). Sensitivity analysis indicated that the two very small treatment‐comparisons Stemmler 2011 Aand Stemmler 2011 B had very little influence on the original OTRR pooled estimate (RR 1.40, 95% CI 1.22 to 1.59 when including Stemmler 2011 Aand Stemmler 2011 B and RR 1.39, 95% CI 1.21 to 1.59 after excluding Stemmler 2011 Aand Stemmler 2011 B) (Analysis 11.3). With regard to Han 2018 A and Han 2018 B, sensitivity analysis indicated that these two treatment‐comparisons had PFS/TTP effect estimates more favourable to platinum than those of the other six treatment‐comparisons in the PFS/TTP meta‐analyses (P = 0.001; Analysis 11.2). Despite this, the PFS/TTP pooled estimate did not change appreciably when Han 2018 A and Han 2018 B were excluded (HR 0.77, 95% CI 0.68 to 0.88 when including Han 2018 A and Han 2018 B and HR 0.82, 95% CI 0.72 to 0.94 after excluding Han 2018 A and Han 2018 B). These results suggest our main findings were not appreciably affected by the inclusion of trials in which the analysis of mTNBC patients was part of a subgroup analysis.
Removal of the trial that used a drug which is now widely considered to be ineffective in the treatment of breast cancer (Carey 2012) had little effect on the OS, PFS/TTP and OTRR point estimates of effect (Analysis 12.1; Analysis 12.2; Analysis 12.3).

Discussion

Summary of main results

Consistent with the findings of the mTNBC subgroup analysis in our previous review (Egger 2017), we found in our current review a marginal OS benefit from platinum‐containing regimens compared to non‐platinum regimens. Specifically, data from six treatment‐comparisons included in the OS meta‐analysis showed a 15% reduction in the risk of death for recipients of platinum‐containing regimens compared to recipients of non‐platinum regimens (P = 0.05). In absolute terms, this 15% risk reduction corresponded to about 55 fewer deaths at one year after metastatic diagnosis for every 1000 mTNBC participants who received platinum‐containing chemotherapy, and about 59 fewer deaths at two years (summary of findings Table 1). Supporting the observed benefit from platinum in terms of OS, platinum‐containing regimens also reduced the risk of death and/or progression (PFS/TTP) by about 23% (P < 0.0001) and increased the likelihood of achieving a complete or partial response (OTRR) by about 40% (P < 0.00001). Moreover, sensitivity analyses suggested that the absence of some treatment‐comparisons from the OS meta‐analysis may have led to an underestimate of the benefit of platinum regimens in terms of OS. While we found a number of statistically significant subgroup differences for OTRR and PFS/TTP (see Effects of interventions), it is difficult to judge the importance or reliability of these findings, given that similar differences were not observed in relation to OS.

It is worth noting that the largest trial, Tutt 2018 did not find an OS, PFS or OTRR advantage for women with mTNBC receiving carboplatin (versus docetaxel). However, in subgroup analysis restricted to 43 BRCA1/2 positive participants, Tutt 2018 found carboplatin was associated with significantly better OTRRs and PFS/TTP. Of the 43 BRCA1/2 positive participants, only 14 (33%) had TNBC, and the remaining 29 (66%) were ER positive, PgR positive and/or HER2 positive (the more common breast cancer clinical subtypes). This suggests that the apparent benefits of platinum for mTNBC patients might be, at least in part, due to the presence of BRCA1/2 mutations in many mTNBC patients, rather than due to the triple‐negative subtype. Furthermore, in the only included trial where 100% of participants were BRCA1/2 positive (Han 2018 A/Han 2018 B), subgroup analysis indicated that the PFS benefits from the two platinum regimens were similar between mTNBC patients and non‐mTNBC patients. Taken together, the findings of Tutt 2018 and Han 2018 A/Han 2018 B and the significant subgroup differences according to BRCA1/2 germline mutation status for PFS/TTP and OTRR (which also included results from Zhang 2018) highlight the importance of performing BRCA1/2 subgroup analysis in future trials assessing platinum chemotherapies for mTNBC patients.

In addition to the BRCA1/2 findings of Tutt 2018, the trial is noteworthy in that it was the largest in this review, it did not find an OS advantage for women with mTNBC receiving carboplatin, and the platinum group performed worst (relative to control) among all trials in this review in terms of PFS and OTRR. Moreover, much of the observed heterogeneity for the PFS and OTRR outcomes appears to be driven by the estimates from Tutt 2018. While it is not clear why the platinum arm faired relatively poorly in Tutt 2018, the trial was the only one to use single agent platinum (carboplatin) in the intervention arm and the only trial to use single agent docetaxel in the control arm.

Another subgroup finding of interest was that the relative benefits of platinum in terms of PFS/TTP and OTRR were greater for trials with 'second‐ or third‐line therapy for ≥ 20% of patients' than for trials with 'first‐line therapy for > 80% of patients' (although this subgroup difference was not observed for the OS outcome). This finding suggests that platinum based regimens may be an effective treatment option following first‐line therapy for mTNBC.

Assessments of toxicity showed that women receiving platinum‐containing regimens experienced higher rates of grade 3 and 4 nausea/vomiting and anaemia than women receiving non‐platinum regimens, but no differences between treatment groups in terms of treatment related death, or grade 3 and 4 hair loss, leukopaenia and treatment discontinuation due to adverse event. In general, however, relatively few intervention‐comparisons could be included in meta‐analyses for adverse events. Better evidence of the toxicity of platinum regimens compared to non‐platinum regimens can be found in our previous analysis (Egger 2017) of adverse events in women with metastatic breast cancer unselected for TNBC. In this regard, our previous review indicated that women receiving platinum‐containing regimens experienced higher rates of grade 3 and 4 nausea/vomiting, anaemia and leukopaenia. More specifically, the higher rate of grade 3 and 4 nausea/vomiting was associated with cisplatin but not carboplatin use, and the increased risk of grade 3 and 4 anaemia was higher for cisplatin recipients than for carboplatin recipients. However, it is likely that newer antiemetics might now lessen the severity of cisplatin‐associated nausea and vomiting. In addition, our previous review also reported that women receiving platinum‐containing regimens experienced a higher rate of grade 3 and 4 hair loss than women receiving non‐platinum regimens. This finding, however, probably relates more to the partner chemotherapy agent than to the platinum itself, as single agent platinum chemotherapies tend not to cause much hair loss.

Overall completeness and applicability of evidence

This review includes data from 10 studies relating to 13 treatment‐comparisons, with publications years ranging from 2005 to 2018. Of the 13 treatment‐comparisons, six (46%), eight (62%) and 10 (77%) provided sufficient data to be included in OS, PFS/TTP and OTRR meta‐analyses, respectively. In general, the number of treatment‐comparisons that had sufficient data on adverse events for mTNBC patients was low, including treatment‐related death (n = 5; 38%), nausea/vomiting (n = 3; 75%), nephrotoxicity (n = 0; 0%), anaemia (n = 5; 38%), hair loss (n = 2; 15%), leukopaenia (n = 5; 38%) and treatment discontinuation due to adverse events (n = 5; 38%). The evidence relating to treatment effects on QoL was wholly incomplete, with no studies reporting QoL results for mTNBC patients.

Although data for the most important outcome (OS) could be included in meta‐analysis for 46% of treatment‐comparisons, the evidence would clearly be more complete if OS data were available for all treatment‐comparisons. Nonetheless, it is somewhat reassuring that in sensitivity analyses, the treatment‐comparisons with PFS/TTP or OTRR data that were not included in the OS meta‐analysis tended to show a greater benefit to platinum in terms of PFS/TTP and/or OTRR results than the six treatment‐comparisons included in the OS meta‐analysis. This provides some evidence that the 'overall' pooled effect estimate for OS was unlikely to show less benefit to platinum if OS data had been available for all 12 treatment‐comparisons that were included in one or more meta‐analyses.

The evidence in this review appears to be generally applicable to the current practice of the treatment of mTNBC for a number of reasons. First, the platinum and non‐platinum regimens used in the included trials contained commonly used chemotherapy drugs currently used in clinical practice to treat metastatic breast cancer including cyclophosphamide, methotrexate and various taxanes. On the other hand, the applicability of the evidence is somewhat reduced by the fact that no trials included an anthracycline (a commonly used class of drug for the treatment mTNBC) in their non‐platinum regimens. Second, the review included trials of women receiving first‐line treatment and women receiving treatment after failure of previous anthracycline or taxane regimens. Third, the trials in the review used the two most commonly used platinum agents for treating metastatic breast cancer, carboplatin and cisplatin. Fourth, this review and our previous review (Egger 2017) are, to date, the only reviews to use meta‐analysis to synthesise the evidence from RCTs assessing whether platinum‐based chemotherapies improved OS for mTNBC participants.

Quality of the evidence

We downgraded the quality of the evidence for the OS effect estimate by one level resulting in moderate‐quality evidence. This was due to imprecision (the CI for the pooled estimate was wide and close to the null) (summary of findings Table 1). This rating of moderate‐quality evidence for OS has improved from low‐quality evidence in our previous review (Egger 2017) because the largest study (Tutt 2018) has now published previously unpublished OS results. This lowered the risk of publication bias for the OS pooled estimate. The quality of the evidence ratings for PFS/TTP and OTRR were downgraded one level for indirectness (because the outcome is a surrogate endpoint) and one level for inconsistency (because there was substantial evidence of heterogeneity). The rating for PFS/TTP was further downgraded one level for suspected publication bias (forest plot asymmetry). As a consequence, we judged the quality of the evidence to be very low for PFS/TTP and low for OTRR.

We graded the quality of evidence for treatment effect estimates of seven key toxicity outcomes (summary of findings Table 2). We judged the quality of evidence to be low for treatment‐related death, nausea/vomiting, anaemia and hair loss. Evidence quality was moderate for leukopaenia and treatment discontinuation due to adverse events. As mentioned above, however, because few treatment‐comparisons could be included in meta‐analyses for adverse events, better evidence of the toxicity of platinum regimens compared to non‐platinum regimens should be found in our previous analysis (Egger 2017) of adverse events in women with metastatic breast cancer unselected for TNBC. However, many studies in that analysis were older trials.

Potential biases in the review process

There were a number of potential biases in the review process. First, it is possible that we may not have identified every eligible study with published results, study protocol or clinical trial registration. This seems unlikely, however, given our highly sensitive search strategies, including access to the Cochrane Breast Cancer Specialised Register maintained by the Cochrane Breast Cancer Group. Second, as with all systematic reviews of clinical trials, there is a risk of reporting bias arising from completed trials that never published their (largely) negative findings (i.e. publication bias). In this review, however, all but one of the included studies were conducted in an era when non‐publication of negative findings was less likely (i.e. due to increasing pressures to pre‐register clinical trials and publish results within reasonable time‐frames). Third, it is possible that trials not specifically assessing mTNBC patients might be more inclined to publish significant mTNBC results from a subgroup analysis that was not pre‐specified in the trial protocol or trial registration. In this review, however, sensitivity analyses indicated that our main results were not appreciably affected by the inclusion of trials in which the analysis of mTNBC patients was part of a post‐hoc subgroup analysis.

Agreements and disagreements with other studies or reviews

Included in the mTNBC subgroup meta‐analyses of our previous review (Egger 2017) were three treatment‐comparisons of 391 women for OS, three treatment‐comparisons of 391 women for PFS/TTP and five treatment‐comparisons of 878 women for OTRR. In the current review, these numbers increased to six treatment‐comparisons of 958 women for OS, eight treatment‐comparisons of 1077 women for PFS/TTP and 10 treatment‐comparisons of 1205 women for OTRR.

In our previous review, we found that for women with mTNBC, platinum‐containing regimens provided benefits in terms of OS (HR 0.75, 95% CI 0.57 to 1.00; low‐quality evidence), PFS/TTP (HR 0.59, 95% CI 0.49 to 0.72; low‐quality evidence) and OTRR (RR 1.33, 95% CI 1.13 to 1.56; low‐quality evidence). In the current review, we found similar, but smaller, benefits for platinum recipients in terms of OS (HR 0.85, 95% CI 0.73 to 1.00; moderate‐quality evidence) and PFS/TTP (HR 0.77, 95% CI 0.68 to 0.88; 1077 women; very low‐quality evidence), and similar, but slightly larger, benefits for platinum recipients in terms of OTRR (RR 1.40, 95% CI 1.22 to 1.59; low‐quality evidence). As a general statement, the findings in this review are similar to the mTNBC‐specific findings in our previous review.

To our knowledge, there is only one other systematic review of randomised trials comparing the effects of platinum and non‐platinum‐containing regimens among participants with mTNBC (Guan 2015). That review, however, only performed meta‐analyses of tumour response rates and not time‐to‐event outcomes. The OTRR meta‐analysis in Guan 2015 comprised three of the 10 mTNBC treatment‐comparisons included in the current review (Bhattacharyya 2009; Carey 2012; Fan 2012). The inclusion of seven additional treatment‐comparisons in the current review resulted in a pooled OTRR estimate of effect (RR 1.40, 95% CI 1.22 to 1.59) that is significantly lower than that of Guan 2015 (RR 2.42, 95% CI 1.66 to 3.53).

While another systematic review also found significant OS and PFS benefits for mTNBC patients who received platinum (Kaya 2018), this review included observational studies and did report RCT‐specific results.

Figure 1

Review 2020: study flow diagram.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 3

Figure 4

Figure 5

Figure 6

Forest plot of comparison: 1 Platinum vs non‐platinum regimens, outcome: 1.1 Overall survival.

Figure 7

Forest plot of comparison: 1 Platinum vs non‐platinum regimens, outcome: 1.2 Progression‐free survival/time to progression.

Figure 8

Forest plot of comparison: 1 Platinum vs non‐platinum regimens, outcome: 1.3 Objective tumour response rate (assessable participants).

Analysis 1.1

Comparison 1: Platinum vs non‐platinum regimens, Outcome 1: Overall survival

Analysis 1.2

Comparison 1: Platinum vs non‐platinum regimens, Outcome 2: Progression‐free survival/time to progression

Analysis 1.3

Comparison 1: Platinum vs non‐platinum regimens, Outcome 3: Objective tumour response rate (assessable participants)

Analysis 2.1

Comparison 2: Platinum vs non‐platinum regimens (subgroup analysis 1: by type of regimen comparison), Outcome 1: Overall survival

Analysis 2.2

Comparison 2: Platinum vs non‐platinum regimens (subgroup analysis 1: by type of regimen comparison), Outcome 2: Progression‐free survival/time to progression

Analysis 2.3

Comparison 2: Platinum vs non‐platinum regimens (subgroup analysis 1: by type of regimen comparison), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 3.1

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 1: Overall survival

Analysis 3.2

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 2: Progression‐free survival/time to progression

Analysis 3.3

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 3.4

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 4: Treatment‐related death (safety population)

Analysis 3.5

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 5: Nausea/vomiting (safety population)

Analysis 3.6

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 6: Anaemia (safety population)

Analysis 3.7

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 7: Hair loss (safety population)

Analysis 3.8

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 8: Leukopenia (safety population)

Analysis 3.9

Comparison 3: Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm), Outcome 9: Treatment discontinuation due to adverse event (safety population)

Analysis 4.1

Comparison 4: Platinum vs non‐platinum regimens (subgroup analysis 3: by first‐line therapy), Outcome 1: Overall survival

Analysis 4.2

Comparison 4: Platinum vs non‐platinum regimens (subgroup analysis 3: by first‐line therapy), Outcome 2: Progression‐free survival/time to progression

Analysis 4.3

Comparison 4: Platinum vs non‐platinum regimens (subgroup analysis 3: by first‐line therapy), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 5.1

Comparison 5: Platinum vs non‐platinum regimens (subgroup analysis 4: by taxane in regimens), Outcome 1: Overall survival

Analysis 5.2

Comparison 5: Platinum vs non‐platinum regimens (subgroup analysis 4: by taxane in regimens), Outcome 2: Progression‐free survival/time to progression

Analysis 5.3

Comparison 5: Platinum vs non‐platinum regimens (subgroup analysis 4: by taxane in regimens), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 6.1

Comparison 6: Platinum vs non‐platinum regimens (subgroup analysis 5: BRCA1/2 mutation status), Outcome 1: Overall survival

Analysis 6.2

Comparison 6: Platinum vs non‐platinum regimens (subgroup analysis 5: BRCA1/2 mutation status), Outcome 2: Progression‐free survival/time to progression

Analysis 6.3

Comparison 6: Platinum vs non‐platinum regimens (subgroup analysis 5: BRCA1/2 mutation status), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 7.1

Comparison 7: Platinum vs non‐platinum regimens (subgroup analysis 6: homologous recombination deficient status), Outcome 1: Overall survival

Analysis 7.2

Comparison 7: Platinum vs non‐platinum regimens (subgroup analysis 6: homologous recombination deficient status), Outcome 2: Progression‐free survival/time to progression

Analysis 7.3

Comparison 7: Platinum vs non‐platinum regimens (subgroup analysis 6: homologous recombination deficient status), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 8.1

Comparison 8: Platinum vs non‐platinum regimens (sensitivity analysis 1: included in OS meta‐analysis vs. not included in OS meta‐analysis), Outcome 1: Progression‐free survival/time to progression

Analysis 8.2

Comparison 8: Platinum vs non‐platinum regimens (sensitivity analysis 1: included in OS meta‐analysis vs. not included in OS meta‐analysis), Outcome 2: Objective tumour response rate (assessable participants)

Analysis 9.1

Comparison 9: Platinum vs non‐platinum regimens (sensitivity analysis 2: Progression‐free survival vs. time to progression), Outcome 1: Progression‐free survival vs time to progression

Analysis 10.1

Comparison 10: Platinum vs non‐platinum regimens (sensitivity analysis 3: Analyses 1 repeated but with random‐effects approach), Outcome 1: Overall survival

Analysis 10.2

Comparison 10: Platinum vs non‐platinum regimens (sensitivity analysis 3: Analyses 1 repeated but with random‐effects approach), Outcome 2: Progression‐free survival/time to progression

Analysis 10.3

Comparison 10: Platinum vs non‐platinum regimens (sensitivity analysis 3: Analyses 1 repeated but with random‐effects approach), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 11.1

Comparison 11: Platinum vs non‐platinum regimens (sensitivity analysis 4: mTNBC patients selected for trial vs. a subgroup of trial), Outcome 1: Overall survival

Analysis 11.2

Comparison 11: Platinum vs non‐platinum regimens (sensitivity analysis 4: mTNBC patients selected for trial vs. a subgroup of trial), Outcome 2: Progression‐free survival/time to progression

Analysis 11.3

Comparison 11: Platinum vs non‐platinum regimens (sensitivity analysis 4: mTNBC patients selected for trial vs. a subgroup of trial), Outcome 3: Objective tumour response rate (assessable participants)

Analysis 12.1

Comparison 12: Platinum vs non‐platinum regimens (sensitivity analysis 5: Analyses 1 repeated but with Carey 2012 excluded), Outcome 1: Overall survival

Analysis 12.2

Comparison 12: Platinum vs non‐platinum regimens (sensitivity analysis 5: Analyses 1 repeated but with Carey 2012 excluded), Outcome 2: Progression‐free survival/time to progression

Analysis 12.3

Comparison 12: Platinum vs non‐platinum regimens (sensitivity analysis 5: Analyses 1 repeated but with Carey 2012 excluded), Outcome 3: Objective tumour response rate (assessable participants)

Summary of findings 1. Platinum compared to non‐platinum regimens for metastatic triple‐negative breast cancer: OS, PFS/TTP and OTRR

Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Platinum compared to non‐platinum chemotherapy regimens for women with metastatic triple‐negative breast cancer
Patient or population: women with metastatic triple‐negative breast cancer (mTNBC) Setting: hospital Intervention: platinum Comparison: non‐platinum chemotherapy regimens
Outcomes	Risk with non‐platinum chemotherapy regimens	Risk with platinum containing regimens	Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Overall survival (OS)	1‐year risk of death		HR 0.85 (0.73 to 1.00)	958 (6)	⊕⊕⊕⊝ MODERATE³	Heterogeneity: Chi² = 5.05, df = 5 (P = 0.41); I² 1%
	510 per 1,000 ¹	455 per 1,000 (406 to 510)²
	2‐year risk of death
	711 per 1,000 ¹	652 per 1,000 (596 to 711)²
Progression‐free survival/time to progression (PFS/TTP)	1‐year risk of progression or death		HR 0.77 (0.68 to 0.88)	1077 (8)	⊕⊝⊝⊝ VERY LOW ^{4 5 6}	Heterogeneity: Chi² = 34.78, df = 7 (P < 0.0001); I² 80%
	936 per 1,000 ¹	880 per 1,000 (846 to 911)²
	2‐year risk of progression or death
	970 per 1,000 ¹	933 per 1,000 (908 to 954)²
Objective tumour response rate (OTRR) (assessable participants)	368 per 1,000 ⁷	515 per 1,000 (449 to 585)	RR 1.40 (1.22 to 1.59)	1205 (10)	⊕⊕⊝⊝ LOW ^{4 5}	Heterogeneity: Chi² = 21.44, df = 9 (P = 0.01); I² 58%
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; HR: Hazard ratio;
GRADE Working Group grades of evidence High quality (⊕⊕⊕⊕): We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality (⊕⊕⊕⊝): We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality (⊕⊕⊝⊝): Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality (⊕⊝⊝⊝ or ⊝⊝⊝⊝): We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Each ⊝ symbol represents a downgrading of the quality of evidence one level from the highest level of 'high quality (⊕⊕⊕⊕).'
¹Estimated from the average of non‐platinum group Kaplan‐Meier probabilities from the 3 highest weighted treatment‐comparisons in Analysis 1.1. ²Estimated as 1000*(1‐S(t)^HR) where S(t) is the estimated probability of survival for non‐platinum participants and HR is the pooled hazard ratio (Guyatt 1998) ³Downgraded quality of evidence one level for ’serious imprecision’ because the confidence interval for the pooled estimate is wide and crosses or nearly crosses unity. ⁴Downgraded quality of evidence one level for ’serious indirectness’ because this outcome is a surrogate endpoint of questionable validity for assessing the more important outcome of OS in the context of metastatic breast cancer (Burzykowski 2008). ⁵Downgraded quality of evidence one level for ’serious inconsistency’ because there was substantial evidence of heterogeneity. ⁶Downgraded quality of evidence one level for suspected publication bias (forest plot asymmetry). ⁷Estimated from all 10 mTNBC treatment‐comparisons in the review with OTRR results.

Summary of findings 1. Platinum compared to non‐platinum regimens for metastatic triple‐negative breast cancer: OS, PFS/TTP and OTRR

Summary of findings 2. Platinum‐containing regimens and toxicity profile

Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Platinum compared to non‐platinum chemotherapy regimens for treatment related death, nausea/vomiting, nephrotoxicity, anaemia, hair loss, leukopaeniaand treatment discontinuation due to adverse event
Patient or population: women with metastatic triple‐negative breast cancer (mTNBC) Setting: hospital Intervention: platinum Comparison: non‐platinum chemotherapy regimens
Outcomes	Risk with non‐platinum chemotherapy regimens	Risk with platinum containing regimens	Relative effect (95% CI)	No. of participants (treatment‐ comparisons)	Quality of the evidence (GRADE)	Comments
Treatment‐related death (safety population)	5 per 1000 ¹	5 per 1,000 (1 to 23)	(RR 1.06, 95% CI 0.24 to 4.61)	843 (5)	⊕⊕⊝⊝ LOW ³	Heterogeneity: P = 0.69, I² 0%
Nausea/vomiting* grade 3 or 4 (safety population)	15 per 1,000 ¹	72 per 1,000 (29 to 177)	(RR 4.77, 95% CI 1.93 to 11.81)	655 (3)	⊕⊕⊝⊝ LOW ³	Heterogeneity: P = 0.32, I² 12%
Nephrotoxicity (safety population)	‐	‐	‐	‐	‐	No trials reported this outcome for mTNBC patients.
Anaemia grade 3 or 4 (safety population)	36 per 1,000 ¹	137 per 1,000 (81 to 231)	(RR 3.80, 95% CI 2.25 to 6.42)	843 (5)	⊕⊕⊝⊝ LOW ²⁴	Heterogeneity: P = 0.04, I² 65%
Hair loss (safety population)	3 per 1000 ¹	1 per 1,000 (0 to 24)	(RR 0.33, 95% CI 0.01 to 8.04)	602 (2)	⊕⊕⊝⊝ LOW ³	Heterogeneity not applicable
Leukopaenia (safety population)	155 per 1000 ¹	169 per 1000 (130 to 220)	(RR 1.09, 95% CI 0.84 to 1.42)	843 (5)	⊕⊕⊕⊝ MODERATE ²	Heterogeneity: P = 0.75, I² 0%
Treatment discontinuation due to adverse event (safety population)	93 per 1000 ¹	82 per 1000 (55 to 123)	(RR 0.88, 95% CI 0.59 to 1.32)	843 (5)	⊕⊕⊕⊝ MODERATE ²	Heterogeneity: P = 0.07, I² 57%
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio;
GRADE Working Group grades of evidence High quality (⊕⊕⊕⊕): We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality (⊕⊕⊕⊝): We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality (⊕⊕⊝⊝): Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality (⊕⊝⊝⊝ or ⊝⊝⊝⊝): We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect Each ⊝ symbol represents a downgrading of the quality of evidence one level from the highest level of 'high quality (⊕⊕⊕⊕).'
¹Estimated from all treatment‐comparisons contributing data for pooling for this outcome (including treatment‐comparisons with non‐estimable effects due to no events in either arm). ²Downgraded quality of evidence one level for 'serious imprecision' because the confidence interval for the pooled estimate is wide. ³Downgraded quality of evidence two levels for 'very serious imprecision' because the confidence interval for the pooled estimate is very wide. ⁴Downgraded quality of evidence one level for 'serious inconsistency' because there was evidence of heterogeneity across studies (P < 0.05) *data on vomiting was included if data on nausea/vomiting was reported separately

Summary of findings 2. Platinum‐containing regimens and toxicity profile

Table 1. Common platinum agents

Generic name	Other names
Carboplatin	Blastocarb, Carboplat, Carboplatin Hexal, Carboplatino, Carbosin, Carbosol, Carbotec, CBDCA, Displata, Ercar, Nealorin, Novoplatinum, Paraplat, Paraplatin AQ, Paraplatin, Paraplatine, Platinwas, Ribocarbo
Cisplatin	Abiplatin, Blastolem, Briplatin,CACP, CDDP, cis‐DDP, cis‐diamminedichloridoplatinum, cis‐diamminedichloro platinum (II), cis‐diamminedichloroplatinum, Cis‐dichloroammine Platinum (II), Cismaplat, Cisplatina, cis‐platinous diamine dichloride, cis‐platinum II diamine dichloride, cis‐platinum II, cis‐platinum, Cisplatyl, Citoplatino, Citosin, CPDD, Cysplatyna, DDP, DDP, Lederplatin, Metaplatin, Neoplatin, PDD, Peyrone's Chloride, Peyrone's Salt, Placis, Platamine, Platiblastin, Platiblastin‐S, Platinex, Platinol‐ AQ, Platinol, Platinol‐AQ VHA Plus, Platinol‐AQ, Platinoxan, platinum diamminodichloride, Platiran, Platistin, Platosin
Oxaliplatin	Ai Heng, Aiheng, diaminocyclohexane oxalatoplatinum, oxalatoplatin, oxalatoplatinum, oxaliplatine, Eloxatin, Dacotin, Dacplat, Eloxatine, 1‐OHP, L‐OHP, oxaliplatin medac

Table 1. Common platinum agents

Table 2. Chemotherapeutic Agents (adapted from Table 1.1 in The Chemotherapy Source Book)

Type of Agent	Action	Includes
Agents that damage the DNA template	by alkylation: nitrogen mustards	cyclophosphamide, melphalan, ifosfamide, chlorambucil
	by alkylation: nitrosureas	carmustine (BCNU), lomustine (CCNU)
	by alkylation: other agents	thiotepa, mitomycin C
	by platinum coordination cross‐linking	cisplatin, carboplatin
	antibiotics	doxorubicin, daunorubicin, mitoxantrone, idarubicin, epirubicin, amsacrine
	podophyllotoxins	etoposide, teniposide
	by intercalation	dactinomycin, mithramycin
	by uncertain mechanisms	bleomycin
Spindle poisons	vinca alkaloids	vincristine, vinblastine, vendesine, vinorelbine
	taxanes	taxol, taxotere
Antimetabolites	thymidylate synthase	5‐fluorouracil
	dihydrofolate reductase	methotrexate

Table 2. Chemotherapeutic Agents (adapted from Table 1.1 in The Chemotherapy Source Book)

Table 3. Number of treatment‐comparisons by subgroup and three outcomes

		Outcome
Subgroup	Treatment‐ comparisons N	Overall survival n (% of N)	Progression ‐free survival/time to progression n (% of N)	Objective tumour response rate n (% of N)
Overall:	13	6 (46%)	8 (62%)	10 (77%)
Type of regimen comparison:
Regimen A + platinum agent vs regimen A	2	1 (50%)	1 (50%)	2 (100%)
Regimen A + platinum agent vs regimen B	9	4 (44%)	5 (56%)	7 (78%)
Single agent platinum vs regimen C	2	1 (50%)	2 (100%)	1 (50%)
Type of platinum agent in platinum arm:
Cisplatin in platinum arm	7	2 (29%)	2 (29%)	6 (86%)
Carboplatin in platinum arm	6	4 (67%)	6 (100%)	4 (67%)
First‐line therapy:
First‐line therapy for > 80% of patients	6	5 (83%)	5 (83%)	6 (100%)
Second‐ or third‐line therapy for >=20% of patients	7	1 (14%)	3 (43%)	4 (57%)
Anthracycline in regimens:
No anthracycline in platinum or non‐platinum regimens	13	6 (46%)	8 (62%)	10 (77%)
Taxane in regimens:
No taxane in platinum or non‐platinum regimens	4	1 (25%)	1 (25%)	4 (100%)
Platinum + taxane vs non‐platinum + taxane regimens	2	2 (100%)	2 (100%)	2 (100%)
Platinum + non‐taxane vs non‐platinum + taxane regimens	5	3 (60%)	3 (60%)	4 (80%)
Platinum + taxane vs non‐platinum + non‐taxane regimens	2	(0%)	2 (100%)	(0%)
BRCA1/2 subtype:
Germline BRCA1/2 mutation #	4	1 (25%)	4 (100%)	2 (50%)
Germline BRCA1/2 wild‐type #	2	1 (50%)	2 (100%)	2 (100%)
Homologous recombination deficiency status:
Homologous recombination deficient #	2	1 (50%)	2 (100%)	2 (100%)
Not homologous recombination deficient #	2	1 (50%)	2 (100%)	2 (100%)
^Numbers for each outcome are the number of treatment‐comparison with sufficient data to be included in meta‐analysis for that outcome. # BRCA1/2 subtype and homologous recombination deficiency status were within‐study subgroupings for Tutt 2018 and Zhang 2018; hence Tutt 2018 and Zhang 2018 both contributed to both BRCA1/2 subroups and both homologous recombination deficiency status subgroups.

Table 3. Number of treatment‐comparisons by subgroup and three outcomes

Table 4. Summary of regimens included in the analysis

Trials ID	Arm 1(platinum‐containing)	Arm 2(control)	First‐line therapy for > 80% of participants	Majority participants anthracycline‐naive
Regimen A + platinum vs regimen A
Bhattacharyya 2009	(Endoxan + with 'cisplatinum')	(Endoxan)	N	N
Carey 2012	C + Cb (Cetuximab + carboplatin)	C (Cetuximab with carboplatin added after progression)	N	N
Regimen A + platinum vs regimen B
Fan 2012	TP (docetaxel + cisplatin)	TX (docetaxel + capecitabine)	Y	N
Mustafa 2019	(cisplatin + gemcitabine)	(paclitaxel + gemcitabine)	Y	N
Han 2018 A	PCP (placebo + carboplatin/paclitaxel)	VT (veliparib + temozolomide)	N	Unknown
Han 2018 B	VCP (eliparib + carboplatin/paclitaxel)	VT (veliparib + temozolomide)	N	Unknown
Stemmler 2011 A	GemCis (gemcitabine + cisplatin)	GemVin (gemcitabine + vinorelbine)	N	N
Stemmler 2011 B	GemCis (gemcitabine + cisplatin)	GemCap (gemcitabine + capecitabine)	N	N
Yardley 2018 A	nab‐P/C (nab‐paclitaxel + carboplatin)	nab‐P/G (nab‐paclitaxel + gemcitabine)	Y	N
Yardley 2018 B	G/C (gemcitabine + carboplatin)	nab‐P/G (nab‐paclitaxel + gemcitabine)	Y	N
Zhang 2018	(cisplatin + gemcitabine)	(paclitaxel + gemcitabine)	Y	N
Icli 2005	Etop + Cis (etoposide + cisplatin)	P (paclitaxel)	N	N
Single agent platinum vs regimen C
Tutt 2018	C (carboplatin)	D (docetaxel)	Y	Y

Table 4. Summary of regimens included in the analysis

Table 5. Summary of outcomes for included trials

Trial ID	OS data useable for HR estimation for mTNBC patients¹	MedianOS time for mTNBC patients²	PFS/TTP data useable for HR estimation for mTNBC patients¹	Median PFS/TTP time for mTNBC patients²	Objective tumour response for mTNBC patients	Treatment‐related deaths for mTNBC patients	Grade III & IV Toxicity for mTNBC patients	Analysed³
Regimen A + platinum vs regimen A
Bhattacharyya 2009	NR	Y	NR	Y	Y	NR	NR	126
Carey 2012	Y	Y	Y	NR	Y	NR	Not useable for meta‐analysis	102
Regimen A + platinum vs regimen B
Fan 2012	Y	Y	Y	Y	Y	Y	Nausea/vomiting Anaemia Leukopaenia Treatment‐discontinuation	53
Mustafa 2019	NR	NR	NR	Y	Y	NR	Not useable for meta‐analysis	110
Han 2018 A	NR	NR	Y	NR	NR	NR	NR	60
Han 2018 B	NR	NR	Y	NR	NR	NR	NR	59
Stemmler 2011 A	NR	NR	NR	NR	Y	NR	NR	15
Stemmler 2011 B	NR	NR	NR	NR	Y	NR	NR	21
Yardley 2018 A	Y	Y	Y	Y	Y	Y	Anaemia Leukopaenia Treatment‐discontinuation	95
Yardley 2018 B	Y	Y	Y	Y	Y	Y	Anaemia Leukopaenia Treatment‐discontinuation	96
Zhang 2018	Y	NR	Y	Y	Y	Y	Nausea/vomiting Anaemia Hair loss Leukopaenia Treatment‐discontinuation	236
Icli 2005	NR	Y	NR	Y	N	Y	Nausea/vomiting Anaemia Leukopaenia	0
Single agent platinum vs regimen C
Tutt 2018	Y	Y	Y	Y	Y	Y	Nausea/vomiting Anaemia Hair loss Leukopaenia Treatment‐discontinuation	376
¹Sufficient data reported to estimate a HR for pooling as outlined by Parmar 1998 and Tierney 2007; this includes Kaplan‐Meier curve, HR and standard error/confidence interval or logrank statistics ²Trials that did not explicitly report median time were classified as NR here regardless of estimable median time from Kaplan‐Meier curve ³Analysed numbers represent the maximum numbers of participants in the treatment‐comparison that were included in a meta‐analysis of OS, PFS/TTP or OTRR (assessable participants). DU: deaths unexplained NR: not reported at all or not reported for mTNBC subgroup OS: overall survival PFS: progression‐free survival TTP: time to progression Y: year reported

Table 5. Summary of outcomes for included trials

Comparison 1. Platinum vs non‐platinum regimens

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Overall survival Show forest plot	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]

1.2 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

1.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

Comparison 1. Platinum vs non‐platinum regimens

Comparison 2. Platinum vs non‐platinum regimens (subgroup analysis 1: by type of regimen comparison)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Overall survival Show forest plot	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]

2.1.1 Regimen A + platinum agent vs regimen A	1	102	(Exp[(O‐E) / V], Fixed, 95% CI)	0.72 [0.46, 1.13]
2.1.2 Regimen A + platinum agent vs regimen B	4	480	(Exp[(O‐E) / V], Fixed, 95% CI)	0.87 [0.70, 1.09]
2.1.3 Single agent platinum vs regimen C	1	376	(Exp[(O‐E) / V], Fixed, 95% CI)	0.87 [0.67, 1.13]
2.2 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

2.2.1 Regimen A + platinum agent vs regimen A	1	102	(Exp[(O‐E) / V], Fixed, 95% CI)	0.56 [0.41, 0.77]
2.2.2 Regimen A + platinum agent vs regimen B	5	539	(Exp[(O‐E) / V], Fixed, 95% CI)	0.68 [0.55, 0.82]
2.2.3 Single agent platinum vs regimen C	2	436	(Exp[(O‐E) / V], Fixed, 95% CI)	1.00 [0.83, 1.22]
2.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

2.3.1 Regimen A + platinum agent vs regimen A	2	222	Risk Ratio (M‐H, Fixed, 95% CI)	2.14 [1.42, 3.23]
2.3.2 Regimen A + platinum agent vs regimen B	7	607	Risk Ratio (M‐H, Fixed, 95% CI)	1.51 [1.29, 1.77]
2.3.3 Single agent platinum vs regimen C	1	376	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.69, 1.23]

Comparison 2. Platinum vs non‐platinum regimens (subgroup analysis 1: by type of regimen comparison)

Comparison 3. Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Overall survival Show forest plot	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]

3.1.1 Cisplatin in platinum arm	2	289	(Exp[(O‐E) / V], Fixed, 95% CI)	0.90 [0.69, 1.18]
3.1.2 Carboplatin in platinum arm	4	669	(Exp[(O‐E) / V], Fixed, 95% CI)	0.83 [0.68, 1.00]
3.2 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

3.2.1 Cisplatin in platinum arm	2	289	(Exp[(O‐E) / V], Fixed, 95% CI)	0.69 [0.54, 0.90]
3.2.2 Carboplatin in platinum arm	6	788	(Exp[(O‐E) / V], Fixed, 95% CI)	0.80 [0.69, 0.93]
3.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

3.3.1 Cisplatin in platinum arm	6	552	Risk Ratio (M‐H, Fixed, 95% CI)	1.61 [1.36, 1.89]
3.3.2 Carboplatin in platinum arm	4	653	Risk Ratio (M‐H, Fixed, 95% CI)	1.16 [0.93, 1.44]
3.4 Treatment‐related death (safety population) Show forest plot	5	843	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.24, 4.61]

3.4.1 Cisplatin in platinum arm	2	289	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
3.4.2 Carboplatin in platinum arm	3	554	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.24, 4.61]
3.5 Nausea/vomiting (safety population) Show forest plot	3	655	Risk Ratio (M‐H, Fixed, 95% CI)	4.77 [1.93, 11.81]

3.5.1 Cisplatin in platinum arm	2	289	Risk Ratio (M‐H, Fixed, 95% CI)	10.89 [2.08, 56.90]
3.5.2 Carboplatin in platinum arm	1	366	Risk Ratio (M‐H, Fixed, 95% CI)	2.47 [0.79, 7.74]
3.6 Anaemia (safety population) Show forest plot	5	843	Risk Ratio (M‐H, Fixed, 95% CI)	3.80 [2.25, 6.42]

3.6.1 Cisplatin in platinum arm	2	289	Risk Ratio (M‐H, Fixed, 95% CI)	6.50 [2.86, 14.77]
3.6.2 Carboplatin in platinum arm	3	554	Risk Ratio (M‐H, Fixed, 95% CI)	2.27 [1.11, 4.62]
3.7 Hair loss (safety population) Show forest plot	2	602	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 8.04]

3.7.1 Cisplatin in platinum arm	1	236	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
3.7.2 Carboplatin in platinum arm	1	366	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 8.04]
3.8 Leukopenia (safety population) Show forest plot	5	843	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.84, 1.42]

3.8.1 Cisplatin in platinum arm	2	289	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.79, 1.36]
3.8.2 Carboplatin in platinum arm	3	554	Risk Ratio (M‐H, Fixed, 95% CI)	1.55 [0.58, 4.12]
3.9 Treatment discontinuation due to adverse event (safety population) Show forest plot	5	843	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.59, 1.32]

3.9.1 Cisplatin in platinum arm	2	289	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.05, 5.44]
3.9.2 Carboplatin in platinum arm	3	554	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.60, 1.36]

Comparison 3. Platinum vs non‐platinum regimens (subgroup analysis 2: by type of platinum agent in platinum arm)

Comparison 4. Platinum vs non‐platinum regimens (subgroup analysis 3: by first‐line therapy)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Overall survival Show forest plot	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]

4.1.1 First‐line therapy for > 80% of patients	5	856	(Exp[(O‐E) / V], Fixed, 95% CI)	0.87 [0.73, 1.03]
4.1.2 Second‐ or third‐line therapy for ≥20% of patients	1	102	(Exp[(O‐E) / V], Fixed, 95% CI)	0.72 [0.46, 1.13]
4.2 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

4.2.1 First‐line therapy for > 80% of patients	5	856	(Exp[(O‐E) / V], Fixed, 95% CI)	0.89 [0.77, 1.04]
4.2.2 Second‐ or third‐line therapy for ≥20% of patients	3	221	(Exp[(O‐E) / V], Fixed, 95% CI)	0.50 [0.38, 0.64]
4.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

4.3.1 First‐line therapy for > 80% of patients	6	947	Risk Ratio (M‐H, Fixed, 95% CI)	1.30 [1.13, 1.50]
4.3.2 Second‐ or third‐line therapy for ≥20% of patients	4	258	Risk Ratio (M‐H, Fixed, 95% CI)	2.05 [1.42, 2.96]

Comparison 4. Platinum vs non‐platinum regimens (subgroup analysis 3: by first‐line therapy)

Comparison 5. Platinum vs non‐platinum regimens (subgroup analysis 4: by taxane in regimens)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Overall survival Show forest plot	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]

5.1.1 No taxane in platinum or non‐platinum regimens	1	102	(Exp[(O‐E) / V], Fixed, 95% CI)	0.72 [0.46, 1.13]
5.1.2 Platinum + taxane vs non‐platinum + taxane regimens	2	148	(Exp[(O‐E) / V], Fixed, 95% CI)	0.61 [0.38, 0.96]
5.1.3 Platinum + non‐taxane vs non‐platinum + taxane regimens	3	708	(Exp[(O‐E) / V], Fixed, 95% CI)	0.92 [0.77, 1.10]
5.1.4 Platinum + taxane vs non‐platinum + non‐taxane regimens	0	0	(Exp[(O‐E) / V], Fixed, 95% CI)	Not estimable
5.2 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

5.2.1 No taxane in platinum or non‐platinum regimens	1	102	(Exp[(O‐E) / V], Fixed, 95% CI)	0.56 [0.41, 0.77]
5.2.2 Platinum + taxane vs non‐platinum + taxane regimens	2	148	(Exp[(O‐E) / V], Fixed, 95% CI)	0.46 [0.30, 0.70]
5.2.3 Platinum + non‐taxane vs non‐platinum + taxane regimens	3	708	(Exp[(O‐E) / V], Fixed, 95% CI)	0.98 [0.84, 1.15]
5.2.4 Platinum + taxane vs non‐platinum + non‐taxane regimens	2	119	(Exp[(O‐E) / V], Fixed, 95% CI)	0.37 [0.23, 0.59]
5.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

5.3.1 No taxane in platinum or non‐platinum regimens	4	258	Risk Ratio (M‐H, Fixed, 95% CI)	2.05 [1.42, 2.96]
5.3.2 Platinum + taxane vs non‐platinum + taxane regimens	2	146	Risk Ratio (M‐H, Fixed, 95% CI)	2.23 [1.48, 3.38]
5.3.3 Platinum + non‐taxane vs non‐platinum + taxane regimens	4	801	Risk Ratio (M‐H, Fixed, 95% CI)	1.18 [1.02, 1.38]
5.3.4 Platinum + taxane vs non‐platinum + non‐taxane regimens	0	0	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable

Comparison 5. Platinum vs non‐platinum regimens (subgroup analysis 4: by taxane in regimens)

Comparison 6. Platinum vs non‐platinum regimens (subgroup analysis 5: BRCA1/2 mutation status)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Overall survival Show forest plot	1	316	(Exp[(O‐E) / V], Fixed, 95% CI)	0.97 [0.81, 1.17]

6.1.1 germline BRCA1/2 mutation	1	43	(Exp[(O‐E) / V], Fixed, 95% CI)	1.00 [0.62, 1.62]
6.1.2 germline BRCA1/2 wild‐type	1	273	(Exp[(O‐E) / V], Fixed, 95% CI)	0.96 [0.79, 1.18]
6.2 Progression‐free survival/time to progression Show forest plot	4	567	(Exp[(O‐E) / V], Fixed, 95% CI)	0.91 [0.76, 1.08]

6.2.1 germline BRCA1/2 mutation	4	176	(Exp[(O‐E) / V], Fixed, 95% CI)	0.43 [0.30, 0.62]
6.2.2 germline BRCA1/2 wild‐type	2	391	(Exp[(O‐E) / V], Fixed, 95% CI)	1.14 [0.93, 1.40]
6.3 Objective tumour response rate (assessable participants) Show forest plot	2	508	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.82, 1.27]

6.3.1 germline BRCA1/2 mutation	2	57	Risk Ratio (M‐H, Fixed, 95% CI)	2.09 [1.17, 3.72]
6.3.2 germline BRCA1/2 wild‐type	2	451	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.71, 1.15]

Comparison 6. Platinum vs non‐platinum regimens (subgroup analysis 5: BRCA1/2 mutation status)

Comparison 7. Platinum vs non‐platinum regimens (subgroup analysis 6: homologous recombination deficient status)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Overall survival Show forest plot	1	316	(Exp[(O‐E) / V], Fixed, 95% CI)	0.97 [0.81, 1.17]

7.1.1 homologous recombination deficient	1	43	(Exp[(O‐E) / V], Fixed, 95% CI)	1.00 [0.62, 1.62]
7.1.2 not homologous recombination deficient	1	273	(Exp[(O‐E) / V], Fixed, 95% CI)	0.96 [0.79, 1.18]
7.2 Progression‐free survival/time to progression Show forest plot	2	328	(Exp[(O‐E) / V], Fixed, 95% CI)	0.94 [0.75, 1.18]

7.2.1 homologous recombination deficient	2	149	(Exp[(O‐E) / V], Fixed, 95% CI)	0.78 [0.55, 1.10]
7.2.2 not homologous recombination deficient	2	179	(Exp[(O‐E) / V], Fixed, 95% CI)	1.09 [0.81, 1.48]
7.3 Objective tumour response rate (assessable participants) Show forest plot	2	328	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.71, 1.15]

7.3.1 homologous recombination deficient	2	149	Risk Ratio (M‐H, Fixed, 95% CI)	0.79 [0.56, 1.12]
7.3.2 not homologous recombination deficient	2	179	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.72, 1.44]

Comparison 7. Platinum vs non‐platinum regimens (subgroup analysis 6: homologous recombination deficient status)

Comparison 8. Platinum vs non‐platinum regimens (sensitivity analysis 1: included in OS meta‐analysis vs. not included in OS meta‐analysis)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

8.1.1 Included in OS meta‐analysis	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.82 [0.72, 0.94]
8.1.2 Not included in OS meta‐analysis	2	119	(Exp[(O‐E) / V], Fixed, 95% CI)	0.37 [0.23, 0.59]
8.2 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

8.2.1 Included in OS meta‐analysis	6	933	Risk Ratio (M‐H, Fixed, 95% CI)	1.30 [1.11, 1.52]
8.2.2 Not included in OS meta‐analysis	4	272	Risk Ratio (M‐H, Fixed, 95% CI)	1.70 [1.33, 2.18]

Comparison 8. Platinum vs non‐platinum regimens (sensitivity analysis 1: included in OS meta‐analysis vs. not included in OS meta‐analysis)

Comparison 9. Platinum vs non‐platinum regimens (sensitivity analysis 2: Progression‐free survival vs. time to progression)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Progression‐free survival vs time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

9.1.1 Progression‐free survival	7	975	(Exp[(O‐E) / V], Fixed, 95% CI)	0.83 [0.72, 0.95]
9.1.2 Time to progression	1	102	(Exp[(O‐E) / V], Fixed, 95% CI)	0.56 [0.41, 0.77]

Comparison 9. Platinum vs non‐platinum regimens (sensitivity analysis 2: Progression‐free survival vs. time to progression)

Comparison 10. Platinum vs non‐platinum regimens (sensitivity analysis 3: Analyses 1 repeated but with random‐effects approach)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
10.1 Overall survival Show forest plot	6		HR (IV, Random, 95% CI)	0.85 [0.72, 1.00]

10.2 Progression‐free survival/time to progression Show forest plot	8		HR (IV, Random, 95% CI)	0.62 [0.45, 0.85]

10.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Random, 95% CI)	1.49 [1.17, 1.89]

Comparison 10. Platinum vs non‐platinum regimens (sensitivity analysis 3: Analyses 1 repeated but with random‐effects approach)

Comparison 11. Platinum vs non‐platinum regimens (sensitivity analysis 4: mTNBC patients selected for trial vs. a subgroup of trial)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 Overall survival Show forest plot	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]

11.1.1 Trial designed to specifically assess mTNBC patients	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.85 [0.73, 1.00]
11.1.2 mTNBC patients were part of a subgroup analysis	0	0	(Exp[(O‐E) / V], Fixed, 95% CI)	Not estimable
11.2 Progression‐free survival/time to progression Show forest plot	8	1077	(Exp[(O‐E) / V], Fixed, 95% CI)	0.77 [0.68, 0.88]

11.2.1 Trial designed to specifically assess mTNBC patients	6	958	(Exp[(O‐E) / V], Fixed, 95% CI)	0.82 [0.72, 0.94]
11.2.2 mTNBC patients were part of a subgroup analysis	2	119	(Exp[(O‐E) / V], Fixed, 95% CI)	0.37 [0.23, 0.59]
11.3 Objective tumour response rate (assessable participants) Show forest plot	10	1205	Risk Ratio (M‐H, Fixed, 95% CI)	1.40 [1.22, 1.59]

11.3.1 Trial designed to specifically assess mTNBC patients	8	1169	Risk Ratio (M‐H, Fixed, 95% CI)	1.39 [1.21, 1.59]
11.3.2 mTNBC patients were part of a subgroup analysis	2	36	Risk Ratio (M‐H, Fixed, 95% CI)	1.69 [0.76, 3.77]

Comparison 11. Platinum vs non‐platinum regimens (sensitivity analysis 4: mTNBC patients selected for trial vs. a subgroup of trial)

Comparison 12. Platinum vs non‐platinum regimens (sensitivity analysis 5: Analyses 1 repeated but with Carey 2012 excluded)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
12.1 Overall survival Show forest plot	5	856	(Exp[(O‐E) / V], Fixed, 95% CI)	0.87 [0.73, 1.03]

12.2 Progression‐free survival/time to progression Show forest plot	7	975	(Exp[(O‐E) / V], Fixed, 95% CI)	0.83 [0.72, 0.95]

12.3 Objective tumour response rate (assessable participants) Show forest plot	9	1109	Risk Ratio (M‐H, Fixed, 95% CI)	1.38 [1.21, 1.57]

Comparison 12. Platinum vs non‐platinum regimens (sensitivity analysis 5: Analyses 1 repeated but with Carey 2012 excluded)