Influence of meteorological factors on the onset of primary spontaneous pneumothorax

Abstract

Introduction

Although links between meteorological conditions and primary spontaneous pneumothorax have been proposed, the reports are controversial. The aim of the study is to correlate between climatic changes and the development of this condition.

Materials and methods

A retrospective chart review included all patient presenting with primary spontaneous pneumothorax to King Fahd Hospital, Imam Abdulrahman Bin Faisal University, Alkhobar, Saudi Arabia, from 1 January 2005 to 31 December 2016. Meteorological data were collected from King Abdulaziz airbase station using an online source for the same time interval. The data were analysed to determine differences in weather conditions between days on which primary spontaneous pneumothorax occurred and those in which it did not. Logistic regression model was used to obtain predicted risks for the onset of primary spontaneous pneumothorax with respect to weather conditions.

Result

Two hundred and eighty-nine patients were found to have primary spontaneous pneumothorax in the 281 days included in the study. Among the meteorological parameters, significant differences were found in average temperature and atmospheric pressure difference between day of admission and two days before the admission, between days with primary spontaneous pneumothorax and days without. There was no significant difference in the other meteorological factors between days with primary spontaneous pneumothorax and days without.

Conclusion

Two hundred and eighty-nine patients were found to have primary spontaneous pneumothorax in the 281 days included in the study. Among the meteorological parameters, significant differences were found in average temperature and atmospheric pressure difference between day of admission and two days before the admission, between days with primary spontaneous pneumothorax and days without. There was no significant difference in the other meteorological factors between days with primary spontaneous pneumothorax and days without.

Introduction

Primary spontaneous pneumothorax (PSP) is defined as an accumulation of air in the pleural cavity with no clinically apparent underlying lung disease.¹ The incidence of PSP shows geographical and gender differences from 6.3 to 18 and from 1.2/100 000/year to 6/100 000/year for men and women, respectively.² The condition occurs most commonly between the second and third decade of life and among tall, thin males who are usually smokers.³

PSP occurs due to rupture of bleb or bullae that develop under the surface of visceral pleura. The reason for the rupture of these bullae is not clear. Many studies have suggested environmental factors, diving activity, listening to loud music or riding on roller coasters to be possible triggers.⁴ Precipitating factors include a change in atmospheric pressure, which may explain the observed clustering of PSP. Some studies confirm a positive association between warm windy days with low atmospheric pressure to be the cause of PSP onset.⁵ The overexpansion of the alveoli due to the trapped air may trigger the occurrence of PSP when the outer pressure falls.⁶ The association between meteorological factors and onset of spontaneous PSP has been investigated in few studies. Correlation between the effect of the meteorological factors and the onset of PSP is still not verified.⁷

There has been no study conducted on this topic in our region. This retrospective study aimed to investigate the possible influence of changes in meteorological factors on the onset of PSP in the Eastern Province of Saudi Arabia.

Materials and methods

Patients

This retrospective study was conducted in King Fahd Hospital, Imam Abdulrahman Bin Faisal University, Al Khobar, Saudi Arabia. Al Khobar is located in the Eastern Province of Saudi Arabia as part of Dammam Metropolitan Area (DMA). In 2012, DMA had an estimated population of 4,140,000. Air pollution in DMA is likely to arise from petrochemical, iron and steel, cement, fertiliser industries, airborne dust, volatile organic compounds and sulphuric acid. The weather in Eastern Province is generally hot and humid, being hottest in July and August. Sand storms are not uncommon, especially during summer.

The study involved 306 patients who were admitted with PSP confirmed by chest radiography over a period of 12 years from 1 January 2005 to 31 December 2016. Of these patients, 17 were excluded from our study because of unclear dates of onset, postoperative recurrence, interstitial pneumonitis or chronic obstructive pulmonary disease. The remaining 289 cases, were retrospectively analysed (Table 1). Because of the retrospective nature of the study, we considered the time of admission for the index day (D). Therefore, for days with PSP (group I) and days without PSP (group II), the measured meteorological values of two consecutive days prior to admission (D-1 and D-2) as well as the day of admission were compared.⁸

Table 1.

Patient demography.

Variable	Findings
Patients (n)	289
Sex (n):	286
Male	3
Female
Age, mean years (range)	23.9 (13–60)
Observation period (days)	4382

Meteorological data

Meteorological data were collected from King Abdulaziz airbase station from online source. Measurement were made on a total of 4383 days over the 12-year study period. All study participants resided within 50 km of King Abdulaziz airbase station and thus were assumed to have been subjected to identical atmospheric conditions. Daily measurement of meteorological parameters included the minimum, average and maximum atmospheric pressure (hPa), temperature (degrees C), humidity (%), wind speed (km/h). Calculations were also made to assess the changes in average atmospheric pressure, temperature and humidity between each study day (D) and one and two days before (D-1 and D-2) in days with and without onset of PSP.

Statistical analysis

Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS, ver.20). The data were described as the mean (median) ± SD. Student’s t-test was used for quantitative variables and Chi-square test for categorical variables. A P value of < 0.05 was considered to be significant. Binary logistic regression was conducted to investigate the association between the occurrence of pneumothorax and meteorological variables. The independent variables were entered into the model sequentially depending on their correlation with the outcome variable. Independent variables that inflated the standard error by more than 10% were excluded to maintain the stability of the model.

Results

A total of 289 cases of PSP occurred in 281 of 4383 days (on eight days, two cases of PSP were reported). The sample included 286 males (99%) and 3 females (1%) and the mean age was 23.3 ± 12.8 years (range 13–60 years). A total of 260 patients (90%) presented with a first episode of PSP and 29 (10%) presented with a recurrence.

All patients presenting with PSP to the emergency department were managed by the thoracic surgery service. Ten (3.5%) patients recovered with rest only. Owing to the advantages of video-assisted thoracoscopic surgery (VATS), we attempted to advise first-episode PSP patients to undergo thoracoscopy. The patients’ choice was in the favour of VATS in most cases. Reports studying VATS in the management of a first attack of PSP showed zero mortality, much less recorded morbidity and lower total costs as compared with conservative therapy.^9,10 Forty-one (14.2%) patients recovered with tube thoracostomy and 238 (82.5%) patients underwent VATS. PSP occurred on 281 days (6.1%) of the 4383 study days.

The trend of PSP by year and months during the study period is plotted on a graph. Figure 1 shows the time series plot of the monthly and yearly incidences of PSP between 2005 and 2016. There is no pattern or trend over time. However, the plot by months showed that the occurrence of PSP is higher in the period from March to May and during the month of October. Hence, the incidence shows a seasonal pattern (Fig 2).

Figure 1.

Cases of primary spontaneous pneumothorax by month, 2005–2016

Figure 2.

Time series plot of the monthly incidence of primary spontaneous pneumothorax between 2005 and 2016

Group I showed an increase in the mean differences in average temperature and average atmospheric pressure two days before the day of admission (D and D-2; 0.82 and 2.4, respectively). These differences were statistically significant (P < 0.001 for both variables) when compared with mean differences in average temperature and average atmospheric pressure for group II. Also, when comparing the mean difference in atmospheric pressure between the day of admission and one day before, (D and D-1), there was an increase by 0.2 hPa, which reaches significance at P = 0.034. The mean difference in average temperature did not reach a significant level one day before admission (Table 2). There were no significant differences between group I and group II regarding other values (Table 3).

Table 2.

Mean differences between group I (days with primary spontaneous pneumothorax, PSP) and group II (days without PSP) in average atmospheric pressure, temperature, and humidity at the day of admission (D) and one (D–D-1) and two days (D–D-2) before admission.

Difference in average	Group I (n = 281)	Group II (n = 4102)	P-value
Atmospheric pressure:
D–D-2	2.4 ± 1.68	0.17 ± 2.9	< 0.001
D–D-1	0.12 ± 1.56	–0.08 ± 2.04	0.034
Temperature:
D–D-2	0.77 ± 0.81	–0.05 ± 2.28	< 0.001
D–D-1	–0.1 ± 0.81	–0.02 ± 1.74	0.147
Humidity:
D–D-2	–1.41 ± 12.78	–0.02 ± 13.55	0.078
D–D-1	0.17 ± 9.9	0 ± 9.84	0.771

Table 3.

Comparison between group I (days with primary spontaneous pneumothorax, PSP) and group II (days without PDP) for different meteorological parameters.

Variable	Group I (n = 281)	Group II (n = 4102)	P-value
Temperature (degrees C):
Minimum	21.8 ± 7	21.5 ± 7.4	0.584
Average	25.7 ± 7.1	25.6 ± 8	0.972
Maximum	34.2 ± 8.5	34.3 ± 8.9	0.845
Humidity (%):
Minimum	23.5± 13.9	23.1 ± 20	0.632
Average	49.1 ± 14.1	47.4 ± 17	0.07
Maximum	73.4 ± 19.9	72.4 ± 19.9	0.403
Atmospheric pressure (hPa):
Minimum	1005.9 ± 8.5	1005.9 ± 8.3	0.906
Average	1008.8 ± 7.4	1008.1 ± 8.1	0.149
Maximum	1009.8 ± 7.8	1009.8 ± 8.2	0.983
Wind speed (km/h):
Minimum	2.4 ± 14.1	3.3 ± 11.6	0.295
Average	13.9 ± 5.6	14 ± 5.4	0.611
Maximum	28.9 ± 16.9	28.5 ± 16.8	0.693

Table 4 shows that there was a significant risk of PSP occurrence with the decrease in both average atmospheric pressure and average temperature two days before admission (odds ratio 1.499 and 1.442, respectively).

Table 4.

Factors associated with occurrence of pneumothorax in the logistic regression model.

Variable	P-value	Odds ratio	95% confidence interval
Difference in average atmospheric pressure (D–D-2)	< 0.001	1.499	1.426–1.575
Difference in average temperature (D–D-2)	< 0.001	1.442	1.352–1.539
Difference in average humidity (D–D-1)	0.055	1.013	0.999–1.027
Average humidity (%)	0.577	1.002	0.994–1.010

Discussion

Our research data demonstrated that a drop in atmospheric pressure and an increase in temperature may play a role in PSP onset. The link between PSP and weather changes has been disputed. Some studies identified a correlation between them,^11,12 while others provided opposing results.^13,14 It is possible that a small meteorological change alone may not fully account for an increased occurrence of PSP.¹⁵ The fact that some PSP admissions are clustered suggested the hypothesis that the meteorological conditions (or some other related factors) could play an important role in the mechanism that leads to the development of pneumothorax.

Change in atmospheric pressure is the most investigated meteorological factor as a cause of PSP, although with conflicting results. It is not clear how variation of the atmospheric pressure might contribute to the occurrence of pneumothorax. According to Bulajich and colleagues,¹⁶ when subpleural blebs or bullae lose communication with lower airway (permanently because of structural changes or transiently because of bronchospasm), air inside blebs/bullae the becomes trapped and rapid equilibration of the pressure gradient between blebs/bullae and the surrounding atmosphere becomes impossible. In these conditions, a transpulmonary pressure gradient originated by the lowering of the atmospheric pressure may result in rapid distension of weakened blebs/bullae walls with the subsequent rupture and occurrence of pneumothorax. This increased tension on the wall of a cyst, resulting in its rupture, has been proven elsewhere as a cause of rupture of pulmonary hydatid cyst.¹⁷

In a retrospective study conducted in China on 337 patients with PSP, a strong correlation was found between the occurrence of PSP and fall in the mean atmospheric pressure. The mean atmospheric pressure in PSP days was obviously lower than that of non-PSP days.¹⁵ Bense⁸ found that a fall in atmospheric pressure of at least 10 millibars within 24 hours was followed by a statistically significant increase in the number of hospital admissions two days later. Similarly, the results of Alfino et al.¹⁸ showed that a significantly large and rapid change in atmospheric pressure was associated with higher incidence of PSP clusters (but not single episodes of PSP), regardless the direction of change.

Not only the magnitude of change in atmospheric pressure, but also the number of exposures to those changes has an impact on the occurrence of pneumothorax, as reported by Scott et al.¹⁹ They stated that exposure to four or more ‘unusual’ change in atmospheric pressure in the four days prior to the onset of symptoms, PSP occurrence was significantly more frequent than expected by chance alone. The unusual change in atmospheric pressure was defined as a fall in atmospheric pressure below the fifth percentile or a rise above the 95th percentile in this time frame.

On the other hand, other investigators have argued that sudden changes in atmospheric pressure cannot be considered a relevant factor triggering PSP. Bertolaccini et al.²⁰ stated that those taking lifts or living in hilly areas are exposed to changes in atmospheric pressure equal or more than typical daily pressure variations. Heyndrickx et al.²¹ did not find significant difference in the variation of mean atmospheric pressure and relative humidity, calculated for each day beginning from the day at which symptoms began, the day before, two days before and three days before the first symptoms (D-day, D-1, D-2 and D-3, respectively). Their work did not conclude occurrence of clusters of a few days in which patients were admitted for a first episode of PSP and explained the reported clusters of PSP admissions as publication bias. Morales and colleagues’²² study, conducted in Spain, also found no significant relationship between the development of spontaneous PSP and either changes in atmospheric pressure or repeated exposure to unusual changes in atmospheric pressure.

In this study, there was a statistically significant difference in the average atmospheric pressure one and two days before PSP admission (D – D-1, and D – D-2, respectively) between group I and group II (days with and days without pneumothorax, respectively). Also, logistic regression model identifies Δ(D–D-2) of average atmospheric pressure as a risk factors for PSP admission.

Several studies reported the relationship between temperature change and PSP occurrence. Our study showed a statistically significant increase in the average temperature two days before PSP admission day in group I and this difference was diagnosed as a risk factor for occurrence of PSP (Table 4). Chen et al.²³ investigated the relationship of seasonal changes in the subtropical Asian climate and climatic effects with PSP. They found a higher monthly incidence of PSP among women during sudden changes in the local temperature. Smit et al.²⁴ reported an increase in mean temperature on a PSP day than the day before. Clusters with PSP were also correlated with the differences in mean temperature between the day of the PSP and the day prior to it. However, they thought that the temperature rise was unlikely to be a causative factor of PSP by itself; they hypothesised that coexisting meteorological phenomena might explain the correlation between temperature changes and the occurrence of PSP.

Several coexisting phenomena were reported in the literature. Monto et al.²⁵ reported that PSP is likely to occur in the time period from 12 noon and 6pm with a temperature of 25 degree Celsius and at 26–36 hours after a drop in atmospheric pressure. Other authors referred to strong wind speed and higher temperature as a coexisting factors that influence the development of PSP.^26,27 Bertolaccini et al.²⁰ supported the opinion that temperature variations do not influence PSP events, as people are commonly exposed to abrupt temperature variations that exceed the usual temperature variations caused by meteorological factors. This conclusion is supported by other authors.

Seasonal variations may influence PSP occurrence. Accard and his colleagues demonstrated the highest frequency of PSP in the period from October to March.²⁸ In our study, PSP admissions were higher in the period from March to May and during the month of October. This may be explained by high temperature variations (and hence the atmospheric pressure according to Boyle’s law) during those months in our region. However, Bulajich et al.¹⁶ reported an equal distribution of PSP throughout the year. Chen et al.²³ stated that neither a particular month of the year nor seasons were found to be significantly associated with the incidence of PSP in either gender. Alifano et al.¹⁸ stated that discrepancies among the aforementioned studies may be due to differences in study design, few patients, different countries with different climatic conditions, exhaustivity of the data on pneumothorax and meteorological conditions and the location of meteorological stations with respect to both the hospital and the patients’ residency.

Other meteorological conditions may also play a role in the onset of PSP. Ozenne et al.²⁹ proposed that bronchoconstriction induced by moist air in the airways might be related to the onset of PSP. Episodes of PSP have been linked with thunderstorms, an extreme situation during which air temperature and humidity change over a relatively short time.²⁴ Concentrations of the major air pollutants NO₂ and O₃ claimed to facilitate the occurrence of PSP.³⁰

There are some limitations in this study. First, the lack of detail on clinical histories. It was not possible to investigate the effects of risk factors, such as smoking and condition of the patients, so patients with secondary spontaneous pneumothorax SP were likely to have been included. Second, retrospective analysis (as is the case with our study) has a natural bias towards positive results. Lastly, the relatively small number of patients might lower the statistical power of the study.

Conclusion

This study demonstrated a significant relationship of PSP occurrence with daily variation in temperature and atmospheric pressure. Further studies are required with larger sample size, focusing on different combinations of weather factors and other potential triggering factors.