Introduction

Lynch syndrome (LS), also known as hereditary non-polyposis colorectal cancer (HNPCC), is perhaps the most common inherited colorectal cancer (CRC) syndrome, underlying approximately 2–4 % of all new cases of colorectal cancer. It develops due to a germline mutation inactivation in one of the four mismatch repair (MMR) genes (MLH1, MSH2, MSH6, PMS2) [1]. Identification of the syndrome is important because of the high risk for, -and penetrance and early onset of cancers in mutation carriers. The lifetime risk for CRC is up to 70–80 % without preventive surveillance, and in many cases cancers develop before age of 50 years. In the early 1990 s The International Collaborative Group on HNPCC (ICG-HNPCC) published guidelines for identifying HNPCC families and recommending use of the Amsterdam I-II criteria [2]. Since the availability of molecular genetic methods for the identification of MMR mutations and antibodies enabling immunohistochemical screening of tumours, systematic screening of all CRC patients for LS has been proposed.

Recent studies have shown that the life-time risk for CRC in LS is 60–70 % in males and 50–60 % in females [3, 4]. Colonoscopic surveillance should be initiated by 20–25 years of age and it has been shown to be effective in reducing the risk for CRC [5, 6]. In a Finnish study, colonoscopy at a 3-year interval reduced the occurrence of CRC by 65 % [5], while even greater reduction may be achieved with a shorter follow-up interval [6]. In addition to CRC, several extracolonic cancer types are associated with LS. The risk for carcinoma of the endometrium or ovary is over 50 % in females. Cancer may also develop in many other organs such as the urinary tract, small bowel and brain in cases of LS [7].

Aim of the study

It is well demonstrated that surveillance reduces the risk for CRC in LS, but a small risk for even advanced CRC exists. The life-long risk for CRC in surveillance has been between 6 and 30 %, depending on the interval between colonoscopies [4, 7, 8]. The important question has been whether CRC- related mortality can be avoided.

In the Finnish studies on the long-term results of colonoscopic surveillance, the absence of CRC- related mortality has been demonstrated [4, 9, 10] and no significant difference found in total mortality between mutation positive and mutation negative individuals 10 years after DNA testing [9].

To confirm the previous findings in this study we analyzed the disease history and cause of death of all the tested mutation carriers (Mut+) (some tested postmortem) as well as their mutation-negative relatives (Mut−). We also analyzed the numbers and types of cancers in all the deceased mutation carriers in the Finnish Lynch Syndrome Registry and the influence of each cancer on the deaths. We also analyzed the number of lethal complications in the numerous surgical and endoscopic procedures the mutation carriers had undergone.

Patients and methods

Finnish Lynch syndrome Registry

The Finnish Lynch Syndrome Registry was established in the early 1980 s [11]. Since the characterization of MMR genes, a predisposing mutation has been verified in 179 LS families (MLH1 132, MSH2 36 and MSH6 11) and over 2500 family members at risk have received genetic counseling and DNA testing. Colonoscopic surveillance at a 2- to 3-year interval has been offered to all mutation carriers from the age 20–25 years. For females, biannual gynecological surveillance has been arranged at the age of 35 years. Compliance with surveillance has been very high [12]. Only three participants have requested a prophylactic colectomy [9], while several females have opted for a prophylactic hysterectomy (unpublished data).

By May 2011 the Finnish LS Registry had data on 2215 tested at risk family members. In 1069 of them, predisposing MMR mutation was verified: (MLH1 840, MSH2 170, MSH6 59) male 521 (48.7 %), female 548 (51.3 %). There was no statistical significant difference in sex ratio between Mut+ and Mut− groups, but the mean age was higher in Mut− compared to Mut+ group (54.3 vs. 52.9 years; P = 0.026). All mutation carriers have given their informed consent prior to their inclusion in Finnish Lynch Syndrome Registry. The survival status of these 1069 mutation carriers was confirmed from the Finnish Population Register Centre, cause of death from the Death Certificate Registry and occurrence of cancers in addition to the Finnish LS Registry also from the Finnish Cancer Registry. The cause of death was always based on the patient’s medical death certificate, and in 34 cases an autopsy was performed (in 6 cancer cases and in 28 other disease). The median follow-up time of all the mutation carriers was 6.68 years (range 0–25 years).

We also analyzed the survival status and cause of deaths of 1146 non-carriers in these 179 families; male 528 (46.1 %), female 618 (53.9 %). The survival status of these non-carriers was confirmed by analogy with the mutation carriers. The follow-up time equals with that of mutation carriers because the genetic testing was organized synchronously for both groups (Table 1).

Table 1 The distribution of deaths according to the MMR gene mutated
Full size table

Statistical methods

Results were given as mean (range). The Kaplan–Meier method was used to calculate cumulative overall and cancer-specific death rate from the date of birth to the end of follow-up and the differences between mutation carriers and mutation negative family members were compared with the long-rank test. SPSS statistical software (version 15.0 for Windows; SPSS. Chicago, IL, USA).

Results

The age distribution of the 1069 mutation carriers is shown in Figs. 1 and 2. Of the 1069 mutation carriers, 151 had succumbed, 97 (64.2 %) to cancer, 51 (33.8 %) to other diseases and 3 (2.0 %) to post-operative complications (Table 1). Only 12 (7.7 %) succumbed carriers had not been affected by cancer, mean age at death of these 12 carriers was 60.25 years (range 35–87 years). Of 151 succumbed individuals, 84 (55.6 %) had had multiple cancers: in 40 cases two cancers, in 20 cases three, in 13 cases four, in 5 cases five, in 4 cases six and in one case seven cancers.

Fig. 1
figure 1

The distribution of surviving mutation carriers N = 918

Full size image
Fig. 2
figure 2

The distribution of the deceased mutation carriers N = 151

Full size image

86 (57 %) of the succumbed carriers were male (mean age 61.3 years, range 33–94 years) and 65 (43 %) female (mean age 65.6, range 31–102 years). The mean age of death from cancer was 63.2 years. The cumulative overall death rate of Mut+ group started to increase at the age of 40 years and was significantly higher (p < 0.001) than that of Mut− family members (Fig. 3). The cumulative cancer-specific death rate of Mut+ group was also significantly increased in Mut+ group (p < 0.001) compared to Mut−family members (Fig. 4). The detailed mortality data for each cancer type is presented in Table 2. The mean age of death from non-cancer causes was 68.8 years. This group included 16 cases (8 males, 8 females) of cardiovascular disease with a mean age mortality of 73 years. Cardiovascular diseases aside, no other prominent disease group was observed in the deaths from non-cancer causes.

Fig. 3
figure 3

Cumulative overall death rate

Full size image
Fig. 4
figure 4

Cumulative cancer-specific death rate

Full size image
Table 2 The occurrence and mortality of the most common LS cancers in the Mut+ group
Full size table

CRC was the most common cancer in the LS mutation carriers. At least one CRC was diagnosed in 390 of the 1069 mutation carriers (male 213, female 177, mean age of diagnosis 45.3 years). Of these 390 carriers, 64 had died and 31 of them from CRC. In 20 cases CRC was diagnosed before LS was identified (in ten cases, the genetic test was performed postmortem). In 11 cases, Lynch syndrome was identified or suspected and surveillance organized before diagnosis of CRC. Of these 11 patients, one case can be considered a true interval cancer leading to the death of the patient. In the remaining cases, the cause was improper surveillance due to either patients or physicians.

The second most common cancer was EC (139 cases). Of this group, 36 had died; EC was a cause of death in 7 cases who didn’t participate gynecological surveillance. Occurrence and mortality of the most common cancers in LS are presented in Table 2.

The total number of operations can only be estimated on the basis of the number of separate carcinomas operated on. This number was 909, and the overall operative mortality was 0.3 %. There was no mortality related to the 4242 colonoscopies performed.

Of the 1146 non-carriers, 44 (3.8 %) had succumbed (mean age 66.5 years, range 43–93 years), 11 (25 %) to cancer (mean age 64.9 years, range 45–79) and 33 (75 %) to other diseases (Table 1). The cancer deaths were related to following cancers: breast 4, stomach 3, pancreas 2, urinary tract 2.

Discussion

Although LS has been considered an optimal target for cancer prevention, even a well-organized colonoscopic surveillance includes life-long risk for CRC. The risk for so called interval cancers observed in colonoscopic surveillance increases to 35 % at age 60 [4]. In a recent retrospective analysis, the risk for metachronous CRC- after segmental large bowel resection was even higher (16, 41 and 62 % at 10, 20 and 30 years of surveillance) despite the fact that the great majority of the patients were undergoing 1–2 yearly colonoscopy [13]. In addition to CRC, many extracolonic cancers, for which no effective screening methods exist, may also be lethal in LS.

The Finnish LS Registry has almost 30 years’ experience of surveillance in LS families. We have previously demonstrated in several studies that our surveillance protocol (colonoscopy every 3 years between 20 and 30 years of age and biannually thereafter) reduces the risk for CRC by over 60 %. This leaves a life-long risk for CRC of 20–30 % but no CRC- related mortality, if the mutation carriers have followed the surveillance instructions [4, 9]. This patient-friendly surveillance protocol has obviously facilitated compliance, which has been over 98 % during the last 15 years [12].

To clarify the cancer-related mortality in the Finnish LS families under surveillance, we analyzed the cause of death for all the succumbed in Mut+ and Mut− groups. Cancer was the cause of death in 64 % of the cases in Mut+ group in contrast to 25 % in the Mut− group and there was a highly significant difference in both overall and cancer specific survival, for a benefit of Mut− group. The hospital records were checked so that we could be sure of the correct cause of death. The age distribution and occurrence of cancers among the surviving and succumbed mutation carriers are shown in Figs. 1, and 2. In the surviving carriers, the occurrence of cancer increases by age, from <2 % at age 20–29, over 50 % at age 50–59 and 82 % at age 79. In the deceased group (Fig. 2) only 12 (8 %) of 151 individuals were free of cancer. In spite of the available surveillance methods, contraction of cancer is very probable in mutation carriers in LS. CRC is the most common cancer. Although, CRC had been contracted by 36.5 % (390) of the mutation carriers, CRC was the cause of death in only 7.9 % (31 patients).

The highest mortality was observed in lung (100 %), bilio-pancreatic (80 %) and brain tumors (78 %), which were diagnosed altogether in 34 (3.2 %) of the 1069 mutation carriers. Urothelial, prostate or gastric cancer was diagnosed in 96 (7.2 %) mutation carriers of whom one in five had died from the disease. Endometrial cancer, the second most common cancer, was diagnosed in 139 (13 %) mutation carriers, of whom only 7 (5 %) had died from this cancer. There was no mortality related to carcinoma of the ovary or small bowel, observed in 21 and 9 mutation carriers, respectively.

Multiple cancers are common in Lynch syndrome. Two or more separate cancers were diagnosed in 57 % of the succumbed mutation carriers and in 14 % of the mutation carriers continuing in surveillance.

Colonoscopy is so far the only surveillance method with indisputable evidence its efficacy in the prevention of cancer and cancer-related mortality in LS, and even this method includes a moderate risk for CRC. For extracolonic cancers there are no effective screening methods. Prophylactic hysterectomy and so-ectomy would be justified for female mutation carriers, especially if molecular genetic defects predisposing to endometrial cancer are present [14]. 47 (48 %) of 97 cancer deaths were due to extracolonic cancer other than EC. Chemoprevention is perhaps the most promising tool for diminishing the risk for these cancers in LS in the future [15]. Genetic counseling is an essential component for cancer prevention in LS [16]. The increased risk for cancer other than CRC or EC should also be emphasized in genetic counseling, and mutation carriers should be encouraged to seek help if they suffer from abnormal symptoms. Correspondingly, physicians should be informed about the risk for extracolonic cancers in Lynch syndrome so that they can be proactive in examining mutation carriers for abnormal symptoms and findings like unexplained anemia which is often seen in small bowel cancers.