This study evaluated the acute health effects of a large oil spill on a large urban population. Subjects in the exposed area had a higher occurrence in symptoms than other areas in the vicinity. There was a clear pattern of decreasing symptoms with increasing distance from the incident site. Symptoms involving eyes, throat, skin, headaches and general malaise were some of the commonly reported symptoms.
Knowledge on the health effects of hydrocarbon exposures mainly comes from studies conducted in occupational settings. Contact with petroleum or exposure to its components may result in skin irritation and stinging or redness of eyes, sore throat, breathlessness, nausea, vomiting headache, dizziness or drowsiness [10]. These are the similar symptoms as found in present study. Major hydrocarbons of toxicological interest are benzene, xylene and toulene. Moderately high concentration of these chemicals irritate mucus membranes particularly the eyes [11]. Sulfur containing components of crude oil may cause respiratory difficulty [12].
High environmental level of pollutants and increased occurrence of headache nausea vomiting, eyes symptoms, respiratory difficulties among the exposed as compared to the control groups suggests the role of pollutants in occurrence of these symptoms.
The proportion of those who developed wheezing with shortness of breath was higher among the exposed as compared to the control groups. Previous studies conducted to assess the health effects of oil spill reported an increase in occurrence of sore eyes, sore throat, headache, skin itching and rash, nausea, vomiting and breathing difficulties among those exposed to the vapors of crude oil [7–9, 13]. The results of our study are consistent with these results; exposed subjects had increased frequency of sore eyes, sore throat, headache, nausea/vomiting, as compared to the non-exposed group. Breathing difficulties also increased among the exposed group. Before oils pill the proportion of those who ever had wheezing with shortness of breath was not different among the exposed and the control groups. Further, wheezing accompanied with shortness of breath is an acute condition that occurs in bouts after exposure to certain substances which trigger it off. These substances in this particular casewere the crude oil vapors. This further suggest role of exposure in occurrence of asthmatic symptoms.
Self reported disturbance in daily routine due to symptoms further substantiate evidence about the severity of symptoms that was higher among the exposed as compared to the non-exposed groups.
We used distance to assess the degree of exposure by selecting the exposed group form the houses located on the shoreline, control group A at 2 km from the source of exposure and control group B at about 20 km from the source. The frequency of exposure as well as strength of associations for many symptoms decreased with the increase in distance. The presence of dose response also suggests a relationship between occurrence of symptoms and exposure.
Those who were exposed for a longer period of time like women among the exposed group, were more likely to develop nausea, vomiting, headache, itchy eyes, scratchy throat and sore throat as compared to men. Women spent most of their time at homes as most of them are housewives, while men were away during the day at their jobs, resulting in an increased exposure to women. These symptoms occurred after exposure to hydrocarbons [10, 11]. In other studies those who were exposed for longer period of time experienced more symptoms as compared to those who were exposed for lesser duration [7].
We analyzed data using two different approaches- multiple linear regression using symptoms score as dependent variable and logistic regression using individual symptom as dependent variable. Both of these approaches revealed strong association of exposure with the symptoms. This provides additional strength to evidence of relationship between exposure to oil spill and ill health.
Strengths and limitations
Strengths include the first study on oil spill health effects to be reported from a developing country, the high participation rate, and collection information on confounders and use of analytical procedure to control the effect of confounders.
Oil spills and other environmental disasters have been reported to increase anxiety, distress and depression among the exposed [14, 15]. However this also depends on level of coverage by the media, concern among the population and availability of means that create awareness and concern among the population. The population affected in current spill has higher educational level than rest of Pakistan but strong civil society organizations that create concerns/awareness are not present. Therefore, level of anxiety may be lesser than that have beenreported from developed countries. We did not have a validated questionnaire to capture anxiety. However, we accounted for anxiety using a single question on anxiousness about effect of oil spill on health.
Since information was based on questionnaires administered at one point in time, there is a likelihood of symptoms being present before the spill occurred. This is the problem due to nature of design used in such studies. However, we addressed this possibility by attempting to determine the occurrence of symptoms related to the specific date of incident. The short recall and correlation of symptoms with a specific exposure facilitated the accuracy of the information. Further, there was the potential for recall bias and over-reporting of symptoms by the exposed group because exposed group may become anxious about their health and work. We addressed the potential problem of differential recall by eliciting information on anxiousness about health due to oil spill and adjusting for this in analysis.
Sample size was based on the availability of funds for conducting this investigation rather than statistical considerations. This has resulted in smaller number of participants in the study that is reflected in our results with small numbers for some variables and resulting wide confidence intervals.
We used three different geographic locations to represent distance form the source of exposure to assign exposure status; this is a crude exposure measurement. Environmental levels of pollutants or biomonitoring could have provided better exposure assessment. However, since exposure was only from one source; distance from the source represents adequate proxy for exposure.