Clinical encounter with three cancer patients affected by groundwater contamination at Camp Lejeune: a case series and review of the literature

The first patient is a 37-year-old Caucasian man who had been relatively healthy until he presented with severe fatigue and muscle ache that persisted for several months. He visited his primary care doctor, and a complete blood count (CBC) revealed a high leukocyte count and the presence of peripheral blasts. When he was sent to an emergency room, his total white blood cell (WBC) count was recorded at 20.5 k/mm³, hemoglobin was 11.6 g/dL, and the platelet count was 146 k/mm³. He underwent a bone marrow biopsy that revealed hypercellular marrow with involvement of T lymphoblastic leukemia. Cytogenetic and molecular analysis was positive for del 11q23 (KMT2A), and rearrangement of 14q32 (IGH). BCR-ABL minor fusion transcript (p190) was detected in a qualitative test, but it was not detected in a quantitative test. The discrepancy was due to a very low level of BCR-ABL fusion in the specimen. Hence the mutation was unlikely a pathogenic driver. Fortunately, there was no evidence of central nervous system (CNS) involvement.

The patient was started on the Linker regimen for T cell acute lymphoblastic leukemia (ALL). Following completion of induction chemotherapy (cycle 1A), a bone marrow biopsy was normocellular, with mature tri-lineage hematopoiesis. Repeat next-generation sequencing (NGS) was negative for any genetic mutation, indicating no measurable residual disease (MRD). The patient received cycles 1B, 1C, 2A, and 2B of the Linker regimen, and then underwent an allogeneic stem cell transplant. He tolerated the procedure well, and his leukemia remains in complete remission.

The second patient is a Caucasian male who has been diagnosed with multiple different types of cancer. He was first diagnosed with prostate adenocarcinoma at the age of 68 after presenting with elevated prostate-specific antigen (PSA) of 4.5 ng/mL. An ultrasound-guided biopsy of a 1 cm nodule in the left lobe was positive for clinical-stage T2aN0 prostate adenocarcinoma. The Gleason score was 3 + 4, and the tumor involved 10% of the biopsy specimen. He underwent brachytherapy with the implantation of iodine-125 seeds. And, at age 74, he was diagnosed with colon adenocarcinoma. A large ascending colon mass was found on a computed tomography (CT) scan during a workup for unexplained anemia. Subsequently, he underwent a right hemicolectomy for a high-grade, poorly differentiated pT4bN0 tumor. After surgery, he received adjuvant FOLFOX chemotherapy and has remained colon cancer-free per radiographic imaging and endoscopic surveillance. He was also diagnosed with lung cancer at the age of 75. A series of CT scans detected a slowly but progressively enlarging mass in the right upper lobe of the lung. When it was brought up as a worrisome finding, the size of the tumor was 10 × 13 mm. He declined to undergo biopsy due to the high risks of the procedure, opting instead for a positron emission tomography (PET) scan. This study confirmed the presence of the right upper lobe pulmonary nodule with a measured standardized uptake value (SUV) of 8.8. After an interdisciplinary discussion, the consensus decision was to obliterate the mass with stereotactic body radiation therapy (SBRT). The patient received a protracted course of 12 fractions of radiation at ×400 cGy per dose, and there has been no evidence of recurrent disease since then.

At the age of 78, he was diagnosed with chronic lymphocytic leukemia (CLL). He had intermittent thrombocytopenia for several years, but workup for the lymphoproliferative disease was negative. However, a recent bone marrow biopsy was positive for kappa monotypic B cells, consistent with CLL. Then, the patient developed Evans syndrome with progressive thrombocytopenia and hemolytic anemia. He was initially treated with rituximab plus prednisone. However, without improvement in the cell counts, the treatment was subsequently switched to a single-agent ibrutinib. After the switch of therapy, his blood counts have stabilized.

This patient was stationed at Camp Lejeune from 1962 to 1987 as an enlisted marine. He lived in a trailer park in Camp Johnson, formerly named Camp Knox. He has been an occasional pipe smoker for about 20 years but had no other exposure to toxic chemicals except for the contaminated water supply. He does not drink alcohol. Although an association is less clear, he and his wife had difficulty conceiving additional children for 10 years after the birth of their first child. His other comorbidities include coronary artery disease requiring stent placement, type 2 diabetes, and depression. There is no one else in the family who had cancer as far as he knows.

He lived at Camp Lejeune while serving as an active-duty marine for 2 years from 1967 to 1968. During the first year, he lived in a barrack in the division headquarters, located at Hadnot Point. During his second year, he and his wife moved to a married housing unit in Tarawa Terrace. Following his discharge from the service, he worked in a suture and contact lens manufacturing factory at Johnson & Johnson, but he denied any occupational exposure to chemicals or toxins. He is a never-smoker and drinks alcohol occasionally. He has a medical history of atrial fibrillation, coronary artery disease, obstructive sleep apnea, and interstitial pulmonary fibrosis. His mother, who was a heavy smoker, died of lung cancer, but no one else in his family has had cancer. Timelines of events for each patient are summarized in Table 1.

PCE (also known as “perc”) and TCE are both halogenated solvents widely used for industrial purposes (Fig. 2). They are colorless, noninflammable, and volatile liquids with an ether-like odor. The vapor from PCE and TCE is heavy, and it readily contaminates ambient urban air. These chemicals are used for dry-cleaning, textile processing, and metal degreasing, and as feedstock in the production of chlorinated chemicals [8, 9]. PCE is relatively stable in its natural state, but it decomposes slowly in contact with moisture. Thus, PCE, TCE, and their by-products, DCE and hydrogen chloride, are commonly found as co-contaminants. These molecules can be absorbed by humans via inhalation or ingestion of contaminated vapor or water, whereas dermal absorption is considered a minor route [10]. PCE and TCE have been classified as probable and known human carcinogens, respectively, by the US EPA, and thus the goal is to achieve zero public exposure. However, the MCL was set at 5 ppb (µg/L) for both PCE and TCE because this is the lowest concentration that is reliably detectable [11]. The Agency for Toxic Substances and Disease Registry (ATSDR) has developed the Hazardous Substance Release and Health Effects Database (HazDat) to collect public health data and compile lists of contaminants specific to known hazardous waste sites. In this database, the most frequently found single-agent contaminant in groundwater was TCE (42.4%), and the most frequently found combination of substances was TCE and PCE (23.5%) [12].

In 1976 and 1978, high levels of PCE were detected in two different areas of Rhode Island by the US EPA, and vinyl liners inside the water distribution pipe were identified as the source of contamination. The vinyl-lined pipes were installed in the 1960s to treat water acidity, and the inner surface of the pipe was coated with vinyl toluene resin in which the solvent used was PCE [13]. This discovery prompted an inspection of water pipes in other states of New England, and an investigation by Massachusetts revealed that a considerable length of the problematic pipes was installed in the upper Cape Cod area, including Barnstable, Bourne, Falmouth, Mashpee, and Sandwich. In certain sites in Falmouth, where water flow was low, the level of PCE reached as high as 18,000 ppb [14]. By the early 1980s, either the affected taps were closed or the pipes were flushed to reduce the PCE concentrations below 40 ppb as a remedial action. Several years after the contamination was first identified, a significantly higher incidence of cancers was observed in towns located in upper Cape Cod relative to the entire state, which occurred in breast, colorectal, lung, and hematopoietic organs [15]. Several case–control studies were subsequently published. One study by Paulu et al. calculated estimated doses of exposure for the involved population based on the residence history, water flow characteristics, pipe age, and dimensions. For people whose exposure level was over the 90th percentile, the adjusted odds ratio (OR) for lung cancer (adjusted for sex, age, other chemical exposures, smoking status, and comorbidities) was elevated at 6.2 [95% confidence interval (CI) 1.1–31.6] after a 7-year latency period, and at 19.3 (95% CI 2.6–141.7) after a 9-year latency period [16]. Another study by Aschengrau et al. observed an elevated adjusted OR for leukemia among those whose exposure level was over the 90th percentile (OR 8.33, 95% CI 1.53–45.29 without latency; OR 5.84, 95% CI 1.37–24.91 with latency) [17].

Another incidence of water contamination with PCE and TCE was reported in 2002 in the city of Redlands in San Bernardino County, California. Water sample tests and hydrogeological studies from this area indicated that the PCE contamination in the water-supply wells began as early as 1980, with levels ranging from 5 to 98 ppb. TCE contamination likely occurred a decade earlier, with levels ranging between 0.09 and 97 ppb [18]. In 1991, the affected wells were removed from the water system, and the water was treated to reduce the concentrations of PCE and TCE. One ecological study calculated standardized incidence ratios (SIRs) of different cancer types in the affected area from 1988 to 1998, based on the actual incidence compared with the expected number of cancers. This study found that the incidence of melanoma and uterine cancer was significantly elevated in the Redlands area during those 10 years, with lower ends of a 99% CI above 1.0 (SIR 1.42, 99% CI 1.13–1.77 for melanoma; SIR 1.35, 99% CI 1.06–1.70 for uterine cancer). However, the prudent authors could not rule out possible confounding effects and stated that distinct area-related socioeconomic status could have impacted their level of healthcare access and vigilance to health-related issues [18].

In New Jersey, routine semi-annual water testing for VOCs became mandatory in 1984. Test results in 1984 and 1985 revealed detectable levels of non-trihalomethane (THM) VOCs, namely PCE, TCE, DCE, and 1,1,1-trichloroethane (TCA), in part of the water systems supplying 20% of the state’s population. An ecological study performed in 1990 in the potentially affected area reported an elevated SIR for leukemia in males (SIR 1.53, 95% CI 1.02–2.21). In this analysis, however, the contaminants were measured as a group of chemicals (non-THM VOC) and not individually, and the total level was only moderately elevated between 37 and 72 ppb [19]. A cohort study in 1994 categorized different levels of TCE in New Jersey water systems based on historical monitoring data in 1978–1984 and mandatory testing results in 1984–1985 and examined differences in rates of hematological malignancies. This study confirmed that the rates of non-Hodgkin’s lymphoma and leukemia were elevated in towns where the average level of TCE was 0.1–5.0 ppb or > 5.0 ppb. However, a dose–response relationship was not so evident, likely because the exposure levels were broadly stratified into < 0.1, 0.1–5.0, and > 5.0 ppb. For patients diagnosed with acute lymphocytic leukemia, for example, the age-adjusted ratio was higher for the 0.1–5.0 ppb exposure level than it was for the > 5.0 ppb level. However, when assessing all types of leukemia, there was a tendency for an increased rate with higher levels of exposure. For males, the age-adjusted rate ratio (RR) was 0.85 (95% CI 0.71–1.02) for the exposure level of 0.1–5.0 ppb, which increased to 1.10 (95% CI 0.84–1.43) for the exposure level of > 5.0 ppb. For females, the association was stronger. The age-adjusted RR was 1.13 (95% CI 0.93–1.37) for the exposure level of 0.1–5.0 ppb and 1.43 (95% CI 1.07–1.90) for the exposure level of > 5.0 ppb. Assuming the validity of this positive association, females were more likely to be diagnosed with leukemia than males. This suggests that sex differences may also play a role in the pathogenesis of leukemia following exposure to these chemicals [20].

Based on the knowledge obtained from laboratory research and epidemiological data, the World Health Organization’s (WHO) International Agency for Research on Cancer (IARC) and the US EPA provide hierarchal classification systems for potential carcinogens (Additional file 1: Tables S1, S2). The IARC and the US EPA have classified PCE as a group 2A and a group B carcinogen, respectively, or “probable human carcinogen” [21, 22]. TCE was classified as a group 1 carcinogen, “carcinogenic to humans,” per IARC standards, and a group A, that is, a “known human carcinogen,” by the US EPA standards [22, 23].

Benzene is a hexagonal aromatic hydrocarbon that is volatile and highly flammable (Fig. 3). It evaporates quickly in room air and emits a unique gasoline-like odor. The vapor is heavier than air, and the liquid form floats on water. It is one of the most widely used chemicals in the United States, and functions as a solvent as well as an intermediate substrate in the various chemical production of plastics, synthetic fibers, rubbers, dyes, and drugs. Benzene is produced naturally in volcanoes and forest fires but is more commonly found in crude oil extracts and produced in the oil-refining process [8]. The general public can be exposed to low levels of benzene from automobile exhaust or industrial emissions. Cigarette smoking is also a major source of benzene exposure in the USA. Rarely, leakage from underground gasoline storage tanks or hazardous waste sites can contain benzene and contaminate drinking water, like what probably happened at Camp Lejeune [24]. Based on EPA guidelines, the MCL of benzene is 5 ppb, although the maximum contaminant level goal (MCLG) is zero [11]. In the HazDat database by ATSDR, benzene was the most frequently found air contaminant (6.0%) and the fourth most frequently found water contaminant (25.8%) in hazardous waste sites [12].

The adverse health effects of chronic benzene exposure are well known, especially hematopoietic toxicities. Animal studies have demonstrated that exposure to benzene is associated with pancytopenia and aplastic anemia as well as chromosomal changes [25, 26]. Different types of skin, gastrointestinal epithelium, liver, respiratory tract, and bone marrow tumors were observed in mice after either prolonged ingestion or inhalation of benzene [27]. The carcinogenicity of benzene was also established in human studies on exposed workers. One of the most extensively studied populations is the rubber hydrochloride workers, the so-called Pliofilm cohort. Exposure to benzene in this cohort occurred at rubber film manufacturing plants in Ohio from the 1930s to the 1970s. During the manufacturing process, the natural rubber was dissolved in benzene and was thinly spread on a conveyer belt. The benzene was then evaporated for recycling, and the rubber film was rolled and recovered from the conveyor belt. Rubber production took place at three different sites in Ohio. In one plant, workers were exposed from 1939 to 1976, and in the other two plants, exposure occurred from 1936 until 1965 [28]. Rinsky et al. meticulously estimated the individual risk of exposure based on company personnel records and past environmental measurements [29]. They reviewed job titles and types of work in the rubber hydrochloride plants to determine “exposure classes,” and combined these data with employment periods and the results of past industrial-hygiene measurements to calculate estimated exposure levels. Then they reviewed the vital status of the employees who worked at one of the three plants for at least 1 day between 1940 and 1965. Deaths were captured as an outcome after January 1, 1950, or earlier if the cumulative personal exposure to benzene reached at least 1 ppm-day (1 day of exposure). As a result, a total of 1165 white men were included in the cohort, contributing to 31,612 person-years at risk. Among these, 330 people died, out of which 15 people had died of lymphatic or hematopoietic cancers as of December 31, 1981. Standardized mortality ratios (SMRs) for both leukemia and multiple myeloma were significantly elevated (SMR 3.37, 95% CI 1.54–6.41 for leukemia; SMR 3.09, 95% CI 1.10–10.47 for multiple myeloma). When the cohort was divided into different exposure strata, there was a progressively increasing dose–response relationship for leukemia, but not for multiple myeloma. In an individually matched nested case–control analysis, it was confirmed that both cumulative dose and duration of exposure were higher for those diagnosed with leukemia and myeloma. Paustenbach et al. considered additional factors in the exposure assessment, including short-term, high-level exposures to vapors, background concentrations in the building, and dermal absorption. These additional exposures significantly increased the estimated total absorption, which was 3–5 times as high [30]. In an updated analysis that added 6 years of observation to the Pliofilm cohort, an additional five cases of leukemia were reported, and there were no additional cases of multiple myeloma. While it was suggested that a higher concentration of benzene was required for leukemogenic potential, the more general association between benzene exposure and leukemia was confirmed [31].

In China, a national occupational survey was carried out in 1981 which reported that more than 500,000 workers in 28,808 factories throughout the country were exposed to various levels of benzene between 1979 and 1981. In this report by Yin et al., the investigators found a higher number of acute or chronic benzene poisoning cases among workers of shoemaking factories where benzene was mixed in the adhesive [32]. Subsequently, they monitored the cancer mortality rate among 28,460 workers who worked for at least 6 months in factories where benzene was used for painting, shoe manufacturing, rubber production, or adhesive production. Compared with workers in factories with no evidence of benzene, radiation, or other known carcinogen exposure, SMRs were elevated for leukemia, lung cancer, primary hepatocarcinoma, and stomach cancer [33]. After these reports, the Chinese Academy of Preventive Medicine (CAPM) collaborated with the US National Cancer Institute (NCI) to expand the analysis and consolidate the assessment, employing qualitative evaluation of the exposure. This analysis included 74,828 benzene-exposed workers from 1972 through 1987 in 672 factories in China and a control group of workers from workplaces where benzene was not used during the same period. The time-weighted average of benzene exposure was estimated by local industrial hygienists and occupational health personnel using measurements of historical ambient benzene levels combined with production and process information for the specific job title in each factory. This updated analysis confirmed that the incidence rate of hematological neoplasms was significantly elevated after benzene exposure. For the workers with estimated average exposure of ≥ 25, RR for all hematological neoplasms was 2.8 (95% CI 1.4–5.7). Moreover, the dose–response relationship was evident with the p-value for trend < 0.5, when the average exposure was divided into < 10, 10–24, and ≥ 25 [34].

Human data supporting benzene carcinogenicity after oral absorption are relatively scarce. Nonetheless, an accumulating number of studies report that water contamination by benzene does occur. For example, benzene was detected in service lines supplying drinking water to standing structures after the record California wildfire in 2018 [35]. The risk of water contamination with benzene is associated with human activities as well. Hydraulic fracturing is used at an increasing frequency for the cost-effective extraction of natural gas, but this process requires the injection of a large amount of water containing chemical additives. The risk of benzene exposure occurs when the flowback water returns to the surface, laced with the remaining additives, solvents, and petroleum-derived hydrocarbons. One study found that the benzene level increased to 148 ppb in a water sample obtained from a reservoir near coal-bed methane fracturing well sites in Sullivan County, Indiana [36]. In addition, there was a report of contaminated water on a container ship, attributed to the internal coating of the water tank [37], and there was a detectable level of benzene with other organic pollutants and organophosphorus pesticides in water reservoirs in the Haihe river basin of China [38]. In the District of Columbia, there was a cluster of leukemia-associated deaths that mysteriously occurred among the mechanics working in the Department of Public Works from 1990 to 1992. In a survey conducted to characterize exposure to carcinogens, the mechanics reported that they used petroleum to clean parts and wash their hands. They also admitted to occasionally siphoning petrol by mouth, resulting in possible oral absorption of benzene [39].

Based on the large volume of evidence characterizing benzene as a carcinogen in both animals and humans, IARC categorized benzene as a group 1 carcinogen, “carcinogenic to humans.” Under the US EPA classification, benzene is a “known carcinogen to humans,” group A. Although human data are relatively lacking for oral carcinogenicity, the US EPA has extrapolated cancer risk from the inhalation data to oral absorption [24].