A meta-analysis of randomized controlled trials (RCT) was performed comparing adjuvant trastuzumab treatment for HER2-positive early breast cancer (EBC) to observation. The MEDLINE, EMBASE, CANCERLIT and Cochrane Library databases, and abstracts published in the annual proceedings were systematically searched for evidence. Relevant reports were reviewed by two reviewers independently and the references from these reports were searched for additional trials, using guidelines set by QUOROM statement criteria.

Pooled results from that five randomized trials of adjuvant Trastuzumab showed a significant reduction of mortality (p < 0.00001), recurrence (p < 0.00001), metastases rates (p < 0.00001) and second tumors other than breast cancer (p = 0.007) as compared to no adjuvant Trastuzumab patients. There were more grade III or IV cardiac toxicity after trastuzumab (203/4555 = 4.5%) versus no trastuzumab (86/4562 = 1.8%). The likelihood of cardiac toxicity was 2.45-fold higher (95% CI 1.89 – 3.16) in trastuzumab arms, however that result was associated with heterogeneity. The likelihood of brain metastases was 1.82-fold higher (95% CI 1.16 – 2.85) in patients who received trastuzumab.

Published randomized controlled trials were eligible for this metaanalysis. Published abstracts were included but unpublished studies were not sought. Studies published in any language were also eligible if they fulfilled the inclusion criteria. No authors were contacted for clarification or verification of patient data. All the trastuzumab adjuvant trials enrolled patients with HER-2-positive (immunohistochemistry 3+/fluorescence in situ hybridization positive or chromogenic in situ hybridization positive for FinHer) invasive breast cancer resected by lumpectomy or mastectomy. Patients could have node positive (all trials) or high-risk, node-negative (N9831, HERA, BCIRG 006, Fin her) disease, and all patients were to receive adjuvant chemotherapy and appropriate radiotherapy and hormonal therapy. Patients were required to have no locally advanced or distant disease and no previous or current cardiac disease. Trastuzumab administered i.v. at any dose and for any time of duration in the adjuvant setting. Differences in mortality, recurrence, Metastases, brain metastases, second tumor non breast cancer and cardiac toxicity rates (classified according to New York Heart Association) were collected. Mortality was defined as death from any cause, recurrence was defined as recurrence of breast cancer at any site; Metastases was defined as of the first distant tumor recurrence, ignoring locoregional recurrences and second breast or non breast cancers. Brain metastases were defined as recurrence in CNS. The term "other tumors no breast cancer" was used to define tumors that appeared after the treatment excluding new case of breast cancer or recurrences of that.

Medline and manual searches were done (completed independently and in duplicate) to identify all published (manuscripts and abstracts) randomized controlled trials (RCTs) that comparing adjuvant trastuzumab treatment for HER2-positive early breast cancer to observation. The Medline search was done on PubMed between 1966 and 2006 with no language restrictions, using the search terms "trastuzumab," breast cancer" or "metastases," adjuvant trastuzumab" or "post operative traztusumab," and "her-2 overexpression." The second search was done through CancerLit, and the Cochrane Library to identify randomized trials published between January 1998 and July 2006, using MeSH headings (trastuzumab, adjuvant trastuzumab, her-2 over expression, breast cancer/sc {Secondary}, ex-lode Clinical Trials, clinical trial {publication type}) and text words (breast, cancer, adjuvant:, trastuzumab, trial, and study) without language restrictions. All the searched abstracts were screened for relevance. Manual searches were done by reviewing articles and abstracts cited in the reference lists of identified RCTs, by reviewing the first author's article, abstract file, from reference lists of retrieved papers, textbooks and review articles. Also, abstracts published in the Proceedings of the Annual Meetings of the American Society of Clinical Oncology (through 2005) were systematically searched for evidence relevant to this meta analysis. The selection of studies for inclusion was carried out independently by two of the authors (V-GA and S-LA). Each study was evaluated for quality using the scale of 1 to 5 proposed by Jadad [8]. If reviewers disagreed on the quality scores, discrepancies were identified and a consensus was reached. Trial data abstraction was also done independently and in duplicate, but abstractors were not blinded to the trials' authors or institution. The reviewers were familiar with the trials, and there were only five trials, so blinding the abstracts was not practical. Any discrepancies in data abstraction were examined further and resolved by consensus.

The NSABP B-31 trial [11] compared four cycles of doxorubicin and cyclophosphamide (AC) followed by four cycles of every-3-week paclitaxel (P) (AC.P, arm1) with AC.P plus 52 weeks of trastuzumab (H) beginning with the first cycle of P (arm 2). The N9831 trial [11,13] was a three-arm study that compared four cycles of AC followed by 12 weekly doses of paclitaxel (AC.P, arm A) with AC.P followed by 52 weeks of H beginning after P (arm B) and with AC.P plus 52 weeks of H beginning with the first P cycle (arm C). Because arm 1 and arm 2 of the B-31 trial are similar to arms A and C of the N9831 trial, the studies were amended to include a joint statistical analysis combining arm 1 and arm A for comparison with arm 2 and arm C. Arm B was excluded from the analysis because H was not given concurrently with P. In both trials, only patients with tumors scored as with 3+ or more staining for HER2 by immunohistochemistry (IHC) or gene amplification by fluorescence in situ hybridization (FISH) were eligible. HER2 testing was centrally confirmed or performed in approved reference laboratories. In an intent-to treat analysis with a combined median follow-up of two years (2.4 years for B-31 and 1.5 years for N-9831), there was a highly significant 52 percent reduction in the risk of disease recurrence with sequential trastuzumab (three-year DFS 87 versus 75 percent, HR 0.48), and despite the short follow-up, a 33 percent reduction in the risk of death (three-year OS 94.3 versus 91.7 percent, HR 0.67) [11].

Patients enrolled in the smaller FinHer trial were randomized to three cycles of docetaxel or vinorelbine followed by three cycles of fluorouracil, epirubicin, and cyclophosphamide [14]. The primary aim of that trial was to compare treatment using docetaxel with treatment using vinorelbine. The subset of women with HER-2-positive tumors (n = 232) was further randomized to either receive or not receive trastuzumab for 9 weeks together with the first three cycles of docetaxel or vinorelbine [14]. The primary end point of FinHer was recurrence-free survival; secondary end points included adverse events (AEs), the effect of treatment on LVEF, time to distant recurrence, and overall survival [14]. Within this subgroup, DFS was significantly better among those who received trastuzumab (89 versus 78 percent, p = 0.01) and there was a trend toward better OS (96 versus 90 percent, p = 0.07). The magnitude of those benefits is similar to those seen in the other trastuzumab trials that used one or two years of treatment.

The HERA trial [12] was an international, multicenter, randomized, controlled trial comparing 1 year or 2 years of trastuzumab given every 3 weeks with observation (no trastuzumab) in patients with HER-2/neu-positive early breast cancer who had completed locoregional therapy and at least four cycles of neoadjuvant or adjuvant chemotherapy from a list of approved regimens. Data were available for 1,694 patients assigned to 2 years of treatment with trastuzumab, 1,694 patients assigned to 1 year of trastuzumab, and 1,693 patients assigned to observation. The results of 1 year of trastuzumab versus observation (first planned interim analysis) were recently reported (n = 3,387) [12]. The most recent report included 1698 controls and 1703 patients treated with one year of trastuzumab, who were followed for an average of 24 months [12]. There was a statistically significant 36 percent reduction in disease recurrence (HR 0.64, three-year DFS of 81 versus 74 percent) as well as a significant improvement in overall survival (HR 0.66, 92 versus 90 percent in the trastuzumab and nontrastuzumab groups, respectively). There were twice as many grade 3 or 4 adverse events with trastuzumab (11 versus 6 percent). The only death attributed to cardiac causes was in the control arm. Trastuzumab was discontinued by 72 women (4 percent) because of cardiac problems.

The BCIRG-006 trial [13] was limited to patients whose tumors had evidence of her-2/neu gene amplification using FISH and included both node-positive and node negative patients. This study evaluated the efficacy and safety of three regimens as adjuvant systemic therapy after surgery: (1) doxorubicin and cyclophosphamide followed by trastuzumab plus docetaxel chemotherapy (AC.TH), (2) docetaxel and carboplatin plus trastuzumab (TCH), and (3) AC followed by docetaxel alone (AC.T) as the control arm. The results of a planned interim analysis of 3,222 patients after approximately one third of the required number of relapses had occurred were recently presented [13]. In a preliminary report with 23 month median follow-up, DFS was significantly better in both trastuzumab arms as compared to AC followed by docetaxel (hazard ratio for DFS 0.49 and 0.61 for AC/docetaxel/trastuzumab and TCH, respectively) [13]. There was no significant difference between the two trastuzumab-containing arms. However, the absolute benefit in DFS from an anthracycline plus trastuzumab compared to TCH was 4 percent. There were fewer symptomatic cardiac events and a lower incidence of asymptomatic LVEF decline with TCH compared to either anthracycline group in terms of two endpoints.

Each published randomized trial reported was assessed for quality using the validated scale developed by Jadad et al [8]. All the studies included were randomized. All studies reported the randomization procedure. The quality scores of included studies are summarized in the table of characteristics, Table ​Table11.

Another question observed in our review was the high incidence of brain metastases in patients who received adjuvant Trastuzumab. The causes of these trends are unknown. Bendell and colleagues [20] retrospectively studied 122 women treated with trastuzumab alone or in combination with chemotherapy for Her-2-overexpressing metastatic breast cancer. Based on a median follow-up of 23 months, 34% of patients were diagnosed with CNS metastases, well above historical rates. At the time of diagnosis of CNS metastasis, 50% of patients were responding to therapy or had stable disease. That report was confirmed by the study of Clayton and colleagues [21] which followed 93 metastatic breast cancer patients. Brain metastases occurred in 25% of patients during a median follow-up period of 10.8 months from the start of trastuzumab therapy. One theory suggests that Her-2 overexpression endows tumor cells with increased metastatic aggressiveness to sites such as the lungs and may similarly augment metastatic propensity to the CNS [22,23]. Second, by allowing patients to live longer, trastuzumab may allow micrometastatic brain metastases to become symptomatic as a natural consequence of an extended life span. A nonexclusive, third theory posits that trastuzumab is effective against systemic metastases but relatively ineffective against CNS metastases due to its poor penetration of the blood-brain- barrier. That hypothesis may extend to cytotoxic chemotherapy as well as Trastuzumab [24]. Limited pharmacokinetic data in support of this hypothesis suggest that systemic administration of trastuzumab results in drug levels in the CSF that are 300-fold lower than in the serum [25]. Also, intrathecal administration of 4D5, the murine precursor of trastuzumab, shows efficacy against a human xenograft of Her-2-overexpressing cancer growing in the leptomeninges, suggesting that trastuzumab could be efficacious if it could penetrate the blood-brain- barrier [26]. Based on our data and other studies [20,21] we predict that brain metastases will become more prevalent as greater control over systemic metastases is achieved, particularly with regard to Her-2-positive tumors.

The incidence of cardiac events with trastuzumab in the EBC setting remained at an acceptable level and was similar across the adjuvant trials, with a reported overall incidence that was 0.6%–3.3% higher [11-14]. Our results analyzing 9117 patients submitted to adjuvant trastuzumab by breast cancer suggest that sequencing the combination of trastuzumab and taxanes plus anthracycline-based chemotherapy increases the risk for class III or IV CHF (OR = 2.45, CI 95 1.89–3.16) However, in our analyses that result was associated with heterogeneity using random effect. Probably due to different definitions of cardiac events, evaluations for cardiac safety, analysis of cardiac end points, and duration of follow-up differed among the adjuvant trials and thus make comprehensive cross-trial conclusions difficult, as showed in Table ​Table2.2. Comparisons regarding cardiac safety should therefore be approached with caution. Increasing evidence suggests that, in patients with metastatic breast cancer who experience a significant LVEF decline, management with beta-blockade and angiotensin-converting enzyme inhibition may allow careful reinitiation of trastuzumab therapy after assessment of the risk-benefit ratio [27]. In the B-31 trial, 27 of the 31 patients in the trastuzumab arm have been followed for>6 months after diagnosis of a cardiac event; 26 were asymptomatic at last assessment and 18 remained on cardiac medication [11]. In this way, the cardiac function of all patients who start treatment with trastuzumab should be monitored and with further follow-up, it will become clearer whether sequential or anthracycline-free regimens carry a lower risk of cardiac toxicity with equivalent efficacy. The mechanism of cardiac dysfunction associated with trastuzumab is not clearly understood, although several hypotheses have been proposed. These include the modification of anthracycline-induced cardiotoxicity, immunemediated destruction of cardiomyocytes, the effects on HER2 signaling pathways that are required for the maintenance of normal cardiac contractility, and the dependence on HER2 for myocyte survival, which is then impaired during trastuzumab treatment [27]. There is increasing experimental evidence supporting a direct toxic effect of HER2 blockade on the heart. HER2 signaling appears to play an important role in embryonic cardiac development and cardioprotection, at least in rodents [28,29]. Those data suggest that trastuzumab-related cardiotoxicity is not immune-mediated or due to effects outside the heart, and it does not result solely from the modification of anthracycline-induced cardiac toxicity. In contrast to anthracycline-related cardiotoxicity, trastuzumab-associated toxicity does not appear to be dose related, and it usually responds to standard medical treatment or the discontinuation of trastuzumab [30,31]. Thus, in the adjuvant setting, continued treatment with trastuzumab is contraindicated if there is any evidence of cardiac dysfunction, and monitoring for early evidence of left ventricular dysfunction is important. In contrast, continuation of trastuzumab therapy or resumption of treatment after resolution of cardiac abnormalities may be safe in some women with metastatic breast cancer who develop early evidence of cardiac dysfunction.

The optimal timing of initiation of trastuzumab in patients with early-stage HER-2/neu-positive breast cancer who have completed adjuvant chemotherapy also has not been defined. The HERA trial allowed patients to begin trastuzumab up to 6 months after local treatment and chemotherapy had been completed. Whether trastuzumab would benefit patients who have completed chemotherapy >6 months previously is unknown. However, in patients with HER-2/neu-positive breast cancer treated with AC followed by paclitaxel chemotherapy, the hazard rate for recurrence increased significantly the first year and remained high in the second year after surgery. This would suggest that trastuzumab therapy is beneficial, although this situation has not been tested in randomized clinical trials. While the sequential administration of trastuzumab after chemotherapy appears to be effective, longer follow- up is needed to determine whether simultaneous and sequential administration with other chemotherapy agents are equally effective and whether there is a population of patients in whom chemotherapy is not necessary (patients with estrogen receptor-positive and HER-2/neu-positive disease, patients with negative lymph nodes and tumors < 1 cm) and for whom trastuzumab alone might represent appropriate adjuvant therapy. Patients in the NSABP B-31, N9831, and BCIRG-006 trials received adjuvant radiation upon completion of chemotherapy during maintenance trastuzumab. Clinical trials should be designed to answer those questions.