Introduction

Stomach cancer is a leading cause of mortality worldwide, accounting for some 650,000 deaths annually [1]. In New Zealand, stomach cancer is a particularly important disease for the indigenous Māori population. It was the fourth most frequently diagnosed cancer among Māori males and seventh among Māori females from 1996 to 2001 [2]. In addition, significant inequities in stomach cancer incidence and mortality have been observed between Māori and non-Māori [3]. Throughout the late 1990s, stomach cancer registration rates for Māori were up to five times those of non-Māori, with similar inequities observed in terms of mortality [4–7]. In line with international trends [1], stomach cancer rates for both Māori and non-Māori fell between 1981 and 2004 [3]; however, ethnic inequities remain [3].

Stomach cancer has a poor prognosis [8, 9]. The 5-year survival for patients with stomach cancer in New Zealand is 20 %, compared to a 5-year survival of 60 % for all cancer sites combined [8]. Māori have 25 % poorer survival (age adjusted) than non-Māori once diagnosed [10] and have poorer survival regardless of stage at diagnosis [2].

A number of factors make the management of stomach cancer particularly complex. The last 2 decades have seen considerable changes in both the epidemiology and treatment of stomach cancer internationally [9, 11–16]. Increasing obesity and associated gastroesophageal reflux is increasing the proportion of patients diagnosed with tumours located in the proximal stomach, especially for white males in developed countries [1]. Infection with Helicobacter pylori (H. pylori) has a known strong association with the development of distally located stomach cancer and is in turn associated with poverty and overcrowding, particularly that experienced in childhood [1, 17]. Other risk factors include tobacco use and heavy alcohol consumption [3, 15, 16, 18]; therefore, many patients have significant levels of comorbidity at diagnosis [11]. Early stomach cancer is often asymptomatic [18] or the symptoms are common and non-specific [16, 18]; thus, many patients are diagnosed at an advanced stage [11, 14, 16]. The primary treatment modality for stomach cancer—surgery—is often complex and can be demanding on the patient, technically challenging for the clinician and considerably resource-intensive to the health care system [9, 14, 16, 19]. As a result of these factors, patients diagnosed with stomach cancer have diverse and complex clinical needs [16], which necessitate care from many different professional groups [11, 16]. As of 2013, there have been no New Zealand guidelines to inform and standardise clinical practice.

For these reasons, the examination of stomach cancer management in New Zealand is important. New Zealand has a publicly funded national health system that provides specialist and hospital care to all residents without patient charges. Despite this, evidence suggests that much of the ethnic cancer survival inequity observed in New Zealand is due to differential access to, and through, services across the cancer control continuum [20, 21]. Therefore, we investigated the presentation, management and survival of stomach cancer in a cohort of newly diagnosed Māori and non-Māori New Zealanders.

Methods

Incident cases of stomach cancer diagnosed between 1 January 2006 and 31 December 2008 were identified from the New Zealand Cancer Register (NZCR). Study inclusion criteria were age 25 years or over, no previous diagnosis of stomach cancer and diagnosis with adenocarcinoma (ICD-10-AM: C1 6.0–16.6, 16.8, 16.9) prior to death. New Zealand comprises two main islands; because around 90 % of Māori are normally resident in the North Island of New Zealand [22], the notes review was limited to this area. All eligible Māori patients, along with a randomly sampled equal number of eligible non-Māori patients, were included. Ethnicity was classified on the basis of NZCR data, where patients are classified as Māori if they have self-identified as Māori on any previous health record. All other patients were classified as non-Māori (see Appendix for progression of exclusions).

Clinical data were extracted from patients’ medical records by an oncology nurse (VS) from public and private hospitals, and where necessary from records held by physicians practicing in private. Data were linked via unique patient identifiers (NHIs) to data collections held in cancer centres, as well as the national administrative hospitalisation dataset (National Minimum Dataset). Data from patients’ medical records were recorded on a standardised pro-forma, double-entered into an electronic database, validation checks carried out and discrepancies resolved.

Data collected included: details of patients’ presentation and diagnosis; tumour characteristics (tumour grade, site and stage at diagnosis); offer, receipt and timing of treatment (surgical and oncological); details of surgical treatment (type and place of surgery, type of surgeon and postoperative complications within 30 days); and palliative care (referral and receipt). Cancer stage was classified according to the TNM classification system [23]. Data were also collected on a specified list of comorbid conditions present at the time of diagnosis as well as any other significant comorbid condition identified. The 12 most common comorbid conditions in this study were included in the analysis. Comorbidities were analysed as both individual conditions and a categorised ‘count’ to assess the overall burden of comorbidity at diagnosis.

When describing the place of surgery, hospitals were categorised into main centres, smaller centres or private hospitals [24]. The type of surgeon was defined according to surgeon self-identification in clinic letters or hospital notes. If specialist status was not clear, this was confirmed with District Health Board records. First intervention was defined as the earliest of radiotherapy, chemotherapy, definitive surgery or another surgical intervention such as abdominal paracentesis, gastric or oesophageal stent, or feeding tube insertion. For analysing receipt and timing of definitive surgery and curative chemotherapy, stage IV and the five unstaged patients were excluded.

Age- and sex-standardised rates were calculated by direct standardisation, using the total New Zealand cancer population (2006–2008) as standard. p values were calculated on crude data from Cochrane-Mantel-Haenszel chi-squared tests stratified by age group or by t test in the case of mean age at diagnosis. Survey methods were used to calculate population estimates for the total New Zealand stomach cancer cohort over the time frame of the study. The final Māori and non-Māori samples were weighted to the total eligible Māori and non-Māori stomach cancer populations. Median times between key steps in the treatment pathway were calculated for the total population and for Māori and non-Māori cohorts. Findings were not corrected for multiple testing.

Cox proportional hazards regression models adjusted for age (continuous), sex, stage (I–IV and unstaged), tumour site (proximal, distal, both proximal and distal, other and missing) and comorbidity count (continuous, 0–12) were used to compare mortality hazard ratios (HRs). Patients who died from other causes and those who did not die were censored at date of death or 31 December 2010 respectively. All analyses were performed using SAS version 9.3. This study was given ethical approval by the Multi-Regional Ethics Committee (ref. no. MEC 10/042/EXP).

Results

The NZCR had a total of 1,115 registrations for stomach cancer (ICD codes C1 6.0–16.6, 16.8, 16.9) nationally during the study period, of which 210 were Māori and 893 non-Māori. Twelve patients had missing ethnicity data and were included in the non-Māori cohort, while 278 patients (16 Māori and 262 non-Māori) were excluded as they resided in the South Island. After all exclusion criteria had been applied, there was a final cohort of 335 patients (172 Māori and 163 non-Māori).

Cohort and disease characteristics

Table 1 shows the characteristics of the final study cohort. The average age of Māori patients at diagnosis was 10 years younger than non-Māori. Among female patients, this ethnic difference was greater still (female mean age at diagnosis: Māori 57 years, non-Māori 70 years). Nearly half (46 %) of all patients were diagnosed at stage IV with no difference in the overall distribution of tumour stage between Māori and non-Māori patients (p = 0.31). Māori appeared less likely to have a poorly differentiated cancer, although when missing grade data were removed there was little difference between Māori and non-Māori in this measure (71 % Māori poorly differentiated compared to 72 % non-Māori).

Table 1 Characteristics of all study-eligible patients (stage I–IV)
Full size table

There were significant differences in the distribution of tumour site between Māori and non-Māori patients; compared with non-Māori, Māori had a higher proportion of distal stomach cancers and a lower proportion of proximal and oesophageal-gastric junction stomach cancers (Table 1). This differential distribution remained when the missing site data were removed (age- and sex-standardised rates for distal tumour site: 58 % Māori, 40 % non-Māori). We observed that the highest proportion of distal tumours occurred in Māori women (age-standardised rates of tumour location: Māori females 54 %, Māori males 31 %, non-Māori females 29 %, non-Māori males 19 %), while the highest proportion of proximal tumours occurred in non-Māori males (non-Māori males 42 %, Māori males 36 %, non-Māori females 28 %, Māori females 16 %).

Patient management

There were no significant ethnic differences in receipt of diagnostic or staging procedures, with most receiving gastroscopy (n = 319) and/or computerised tomography (CT) scan (n = 300). Only four received an endoscopic ultrasound.

Of the 172 patients with stage I–III disease, two-thirds had definitive surgery (Table 2). Māori and non-Māori patients had similar rates of definitive surgery; however, when compared with non-Māori patients, Māori were considerably less likely to have surgery performed by a specialist upper gastrointestinal surgeon (38 % for Māori and 79 % non-Māori patients, p < 0.01) and less likely to have surgery in a main centre (44 % for Māori and 87 % non-Māori patients, p < 0.01). Even when stratified by surgery type Māori remained less likely than non-Māori to have surgery performed by a specialist surgeon or to be treated in a main centre. Overall, few patients had surgery in a private facility.

Table 2 Characteristics of definitive surgery for stage I–III patients
Full size table

Of the patients with stage I–III disease, 49 % were referred to medical oncology; however, few patients received chemotherapy in conjunction with surgery (Table 2). Overall, the median waiting time from date of diagnosis till first intervention was 31 days, till definitive surgery 35 days and till referral to medical oncology 25 days. There were no statistically significant differences in waiting times to treatment between Māori and non-Māori patients. Māori (88 %) and non-Māori (83 %) stage IV patients were similarly likely to be referred to a palliative service (palliative chemotherapy, palliative radiotherapy or other palliative care).

A substantial proportion of patients with stage I–III disease appeared to have no treatment at all (19 % of stage I–III patients). Upon investigation, this group were older (mean age 79 years) than the stage I–III patients that did have treatment (mean age 63 years) and were more likely to have a higher number of comorbid conditions at diagnosis (mean comorbidity count of 3.06 versus a mean comorbidity count of 1.55 for those stage I–III patients that did have treatment).

Māori appeared to have poorer cancer-specific survival than non-Māori. After adjusting for age, sex, stage, tumour site and comorbidity, Māori patients were 27 % more likely to die of their stomach cancer than non-Māori although the difference was not statistically significant (HR 1.27, 95 % CI 0.96–1.68).

Discussion

This study found both similarities and differences in the presentation, management and survival of stomach cancer for Māori when compared with non-Māori patients in New Zealand. While there were no significant differences in cancer grade or stage at diagnosis, Māori were younger and presented with a much higher proportion of distal stomach cancers when compared with non-Māori patients. Māori had significantly higher prevalence of comorbid congestive heart failure and renal disease. Of those patients diagnosed with stage I–III disease, Māori were equally likely to receive definitive surgery as non-Māori, although Māori were less likely to have surgery performed by a specialist upper gastrointestinal surgeon and less likely to be treated in a main centre. Few patients overall received chemotherapy. Māori in this cohort appeared less likely to survive once diagnosed with stomach cancer, although the study was underpowered to statistically confirm a 27 % excess mortality among Māori patients.

Our observation of a higher proportion of distal cancer for Māori is in keeping with previous New Zealand-based studies [25, 26]. This finding suggests that there may be differing aetiological factors driving the high incidence rates of stomach cancer observed for Māori. Infection with H. pylori and smoking have both been shown to be more likely to lead to the development of distal stomach cancer over proximal [1, 9, 25, 27–29]. The high proportion of distal stomach cancer among Māori women when compared with non-Māori women may be related to their higher rates of H. pylori in combination with a very high smoking prevalence and younger age at initiation [25, 30–32]. Māori women have one of the highest rates of smoking in the world, more than Māori men and over twice that of non-Māori women [30]. These two factors are thought to interact to increase the risk of stomach cancer more than would be expected given each risk factor alone [29, 33]. Further research into the risk factors of stomach cancer for Māori, and whether there are significant gender differences as suggested by this study, is warranted. These findings add weight to a continued emphasis on reducing smoking as well as the development of interventions to prevent the transmission of (and to treat) H. pylori, particularly among Māori.

Our observation that Māori were more likely to have comorbidities is consistent with previous studies that have found Māori patients to have higher rates of comorbidity than non-Māori patients with cancer [34]. Comorbidity is known to impact on the quality of care received by patients and on the likelihood of survival from cancer [34–37]. The fact that the group of stage I–III patients in this study who did not receive any treatment had higher levels of comorbidity than those who did receive treatment suggests that comorbidity is playing a role in the decision to treat. The effect of comorbidity on treatment and subsequent survival requires further investigation.

Surgery was the primary treatment modality over our study period, which is consistent with international guidelines in use at the time [12, 14, 19]. However, our findings suggest that the guideline recommendations were not being met, particularly for Māori patients. International guidelines, published in 2002 and 2006, recommended that all patients should have treatment planned within the multidisciplinary context and that at all stages of disease surgery should be undertaken by experienced surgeons in high-volume specialised units [14, 19] with appropriate postoperative care available [14]. Māori in this study were more likely to have distal disease and thus more likely to undergo less complex partial gastrectomy. They were also, however, less likely than non-Māori to have a specialist upper gastrointestinal surgeon or to have surgery performed in a main centre with specialist postoperative support whether their operation was a partial gastrectomy or the more complex total gastrectomy. While this may in part be due to more Māori living in minor urban [22] and rural areas [38], these findings indicate differential access to specialised surgical stomach cancer services for Māori.

The only current guideline related to stomach cancer in New Zealand during the time frames of this study (‘Suspected cancer in primary care’) [39] advises primary care practitioners to consider stomach cancer at a younger age (suggesting 10 years earlier) when treating Māori patients compared to the general population. This recommendation is supported by the findings of the current study. The differential average age at diagnosis is likely due to the younger age structure of the Māori population [22], but also may indicate a true younger age at onset possibly due to a higher prevalence of known risk factors among Māori [25, 30–32].

International stomach cancer treatment guidelines have changed over the last decade, particularly in the area of medical oncology. In 2002, the UK guideline recommended that chemotherapy is ‘not standard practice’ but is given in the context of clinical trials only [14]. By 2011 the updated UK guideline advised clinicians that neoadjuvant (before surgery) and adjuvant (after surgery) chemotherapy ‘conveys a significant survival benefit’ and recommended that patients be given adjuvant treatment if neoadjuvant has been missed [15]. Despite increasing evidence during the time frames of this study that medical oncology is an effective treatment modality for stomach cancer, little chemotherapy was given in this study. A New Zealand study published in 2002 [9] found that few patients with stomach cancer received multimodality therapy despite changing evidence at the time to support such treatment. The authors also highlighted the evidence for better staging and treatment planning offered by endoscopic ultrasonography in conjunction with CT scanning [9]. Their study, using data collected between 1995 and 1997, found that 10 % of patients receiving an operation for a gastro–oesphageal tumour had an unnecessary ‘open and close’ surgical procedure, indicating the need for better surgical planning information gained by endoscopic ultrasonography prior to surgery. Despite these recommendations, our findings indicate that clinical practice remained unchanged a decade later, with only four patients receiving an endoscopic ultrasound in the study cohort. The absence of New Zealand-based guidelines to inform clinical practice during the time period covered by this study may be significant. The government has recognised the need for uniform treatment guidelines for stomach cancer and national standards have recently been developed [40].

A key strength of this study is that it is based on a full clinical notes review, which allowed us to collect comprehensive presentation and management data on all eligible patients and conduct a detailed comparison between Māori and non-Māori patients. Importantly, while the NZCR reported 101 patients within this cohort as unstaged, we were able to determine stage at diagnosis for all but five patients and thus include this important treatment and prognostic factor in our analysis. We were however unable to obtain complete data on the key tumour variables of grade and site or on patient smoking status. Firm conclusions are limited by our small sample size, especially considering 46 % (those patients stage IV at diagnosis) were excluded from some analyses. Additionally, data on histological subtype (diffuse or intestinal) were not collected. Māori have been shown to be more likely to present with diffuse stomach cancer, which is thought to negatively impact on prognosis [17, 25]. Finally, it is not possible to rule out the possibility of chance findings, particularly given the number of comparisons. However, the key statistically significant findings (such as Māori having higher levels of comorbidity, more distal cancers and less access to specialised care) are those where we had a priori expectation of finding differences; reducing the possibility that these are chance findings.

Conclusion

The investigation of stomach cancer and its management is a high priority for Māori cancer control. We found evidence of differential presentation, especially tumour site, and in access to specialised surgical services for Māori stomach cancer patients compared to non-Māori. Māori also appear 27 % less likely to survive once diagnosed. These findings support the development and implementation of national stomach cancer treatment standards for New Zealand. They also highlight the imperative that these standards have an equity focus and prioritise the needs of Māori.